Skip to main content

Full text of "Cellular mechanisms for the regulated degradation of aldolase B"

See other formats


Peter P. Susan 












General Concepts of Protein Degradation 1 

Background for Fructose 1,6-Diphosphate Aldolase B 3 

Mechanisms for Degradation Aldolase B 8 

Hypothesis for Stress-Induced Degradation of Aldolase B 28 

General Strategy 30 


Cell Lines and Culturing 33 

Plasmid Vector Construction and Mutagenesis 35 


Expressing Epitope-Tagged Aldolase B in Cell Lines 41 

Immunofluorescence 44 

Antibodies 44 

Viability Assays 48 

Subcellular Fractionation 49 

Enzyme Assays 52 

Protein Analysis 53 

Stress-Induction of Protein Degradation 55 

Protein Degradation 57 



Introduction 59 

In Vivo Multiubiquitination of Aldolase B 60 

Multiubiquitinated Aldolase B is Denatured and Enriched in Lysosomes 68 

Heat Stress-Induced Delivery of Aldolase A to Lysosomes Requires 
Ubiquitination 69 

Heat Stress-Induced Proteolysis of Aldolase B Requires Ubiquitination 75 


Ubiquitin-Mediated Autophagic Degradation Occurs in E36AB Cells 93 



Introductions 101 

Ubiquitin-Independent Cytosolic Proteolysis of Aldolase B 1 02 

Temperature-Dependent Cytosolic Proteolysis in Fao Cells 1 06 

Starvation-Induced Autophagic Degradation of Aldolase B in Fao Cells.... 1 10 

Temperature-Dependent Autophagy and Cytosolic Proteolysis 118 

A Model For the Degradation of Aldolase B 123 


Introduction 127 

Transient Expression of RABM Mutations in Putative Lysosome 
Targeting Signals 130 

Starvation Induces Autophagic Degradation in HuH7 Cells 139 

Transient Expression Does Not Affect Starvation-Induced Degradation 
of RABM 140 

Site-Directed Mutations Did Not Affect Wildtype Activity of RABM 142 

Glutamine Residue #1 1 1 is Required for Starvation-Induced Degradation 
of Aldolase B 143 

Glutamine #111 Specifically Mediates Starvation-Induced Degradation of 
Aldolase B 146 



Introduction 151 

Autophagy and Ubiquitination 151 

Clues from Temperature-Dependent Cytosolic Proteolysis and 
Lysosomal Degradation 154 

Signal-Mediated Targeting 157 

Present and Future Contributions to the Field of Protein 
Turnover 153 



Abstract of Dissertation Presented to the Graduate School 
of the University of Florida in Partial Fulfillment of the 
Requirements for the Degree of Doctor of Philosophy 



Peter P. Susan 

August 1998 

Chairman: William A. Dunn, Jr. 

Major Department: Anatomy and Cell Biology 

Stress-induced degradation of abundant long-lived cytosolic housekeeping 

proteins was examined using liver aldolase B as a model protein. Heat stress increases 

ubiquitination that mediates autophagic degradation of long-lived proteins in E36 

Chinese hamster cells. During starvation, major multiubiquitinated proteins (e.g., Ub68) 

increased in lysosomes (rat liver and Fao hepatoma cells) and were antigenically 

characterized as aldolase B-ubiquitin conjugates. Compared with controls, heat stress 

increased endogenous aldolase A activity in lysosomes of E36 cells by >twofold. Heat 

stress was non-permissive for ubiquitination in E36-derived ts20 mutant cells and failed 

to increase aldolase activity in ts20 lysosomes. Myc-tagged aldolase B (RABM) 

expressed in E36 cells underwent limited proteolytic processing in lysosomes that failed 


to occur in heat stressed ts20 cells. The results suggested that during stress (starvation 
or heat), aldolase A and aldolase B can undergo ubiquitin-mediated autophagic 
degradation. Long-lived protein degradation was a continuous function of temperature, 
indicating heat stress-induced rates were due to thermodynamic stimulation of chemical 
reactivity. Lysosomal inhibitors distinguished proteolysis in lysosomes from that in 
cytosol. Complete autophagic degradation to amino acids in lysosomes was highly 
temperature-dependent compared to a relatively constant rate in cytosol. HA-tagged 
human aldolase B (HAHAB) in Fao cells and RABM in E36 cells underwent proteolysis 
in cytosol that had temperature-dependence paralleling complete degradation of proteins 
in lysosomes. Lysosomal degradation was ubiquitin-dependent (blocked in heat stress 
ts20 cells), but cytosolic proteolysis of RABM was not. Results suggest a possibly 
shared temperature-dependent cytosolic mechanism that limits rates for partial cytosolic 
proteolysis and complete lysosomal degradation of long-lived proteins. Three peptide 
motifs for signal-mediated targeting to lysosomes during starvation occur in aldolase B. 
These were mutated in RABM. Starvation-induced degradation of mutant and wildtype 
RABM expressed in HuH7 human hepatoma cells were measured. Starvation-induced 
degradation of RABM (aldolase B) specifically required a glutamine at residue #111 
suggesting that the corresponding peptide motif, IKXDQ, is a targeting signal 
functionally demonstrated in living cells. Evidence was provided for three previously 
unknown mechanisms for stress-regulated degradation of aldolase B: (1) ubiquitin- 


mediated autophagic degradation in lysosomes, (2) temperature-dependent cytosolic 
proteolysis during heat stress, and (3) signal-mediated degradation during starvation. 



General Concepts of Protein Degradation 

In cells, different proteins have different functions and occur at different levels as 
needed. The functional activities and locations of proteins are regulated to integrate with 
each other, maximizing survival. Proteins can be regulated by a variety of mechanisms, 
but available concentration of each protein fundamentally determines maximal function 
(Doherty and Mayer, 1992). Cells adapt to environmental change by altering amounts of 
different proteins. Some increase, others decrease, and the rest are constant (Doherty 
and Mayer, 1992). Such adaptation of different proteins requires preferential 
mechanisms for regulating synthetic or degradative rates in response to environmental 

Under constant conditions, protein synthesis is independent (zero order 
relationship) of the protein concentration, but degradation is directly proportional (first 
order relationship) to protein concentration (Doherty and Mayer, 1992). Synthesis 
increases protein concentration, causing degradation to increase until synthetic and 
degradative rates are equal. In this way, a balance between synthesis and degradation 
determines the available concentration of a protein. If environmental conditions change, 
then cells can adapt protein concentrations by modulating synthesis or degradation 

(Olson, et al., 1992; Mortimore, 1987). This study examines mechanisms of protein 

degradation that respond to environmental changes. 

Continual synthesis and degradation results in constant turnover of proteins 
which can be described by either the fractional degradation rate or the half-life of the 
protein. The fractional degradation rate of a protein (degradative rate constant), kd, is 
defined as the fraction of the initial protein degraded in a given time. The kd is calculated 
from measurements of labeled protein lost per time. Half-life, ty„ is defined as the time 
for turnover of half the protein. Under equilibrium conditions, kd and ty 2 are constant and 
directly related to each other by t Vl = ln(2)/ kd, allowing calculation of t /2 from 
experimentally determined kd (Doherty and Mayer, 1992). 

Proteins are categorized as short-lived for ty 2 < 1 h or long-lived for t Vx >1 h. For 
short-lived proteins, protein concentrations respond more to changes in synthesis (Olsen, 
et al., 1992). For long-lived proteins, protein concentrations respond more to changes in 
degradation (Olsen, et al., 1992; Mortimore and Poso, 1987). Detailed reasoning for 
this is described elsewhere (Doherty and Mayer, 1992). Many short-lived proteins utilize 
a well-characterized mechanism for degradation by a cytosolic protease complex called 
the proteasome (Rock, et al., 1994; Ciechanover and Schwartz, 1994; Hochstrasser, 
1992). Relative to short-lived proteins, mechanisms for the degradation of long-lived 
proteins are poorly characterized. This dissertation examines mechanisms for stress- 
inducible degradation of a long-lived cytosolic enzyme, fructose 1 ,6-diphosphate 
aldolase (see next section). 

Mechanisms for degradation of proteins become enhanced during environmental 

stress. Increased temperature (Bates, et al., 1982; Hough and Reichsteiner, 1984) or 

nutrient deprivation (Mortimore and Poso, 1987) are known to increase the degradation 

of long-lived proteins. Stress-induced degradative mechanisms are of special interest, 

because they mediate regulated changes and themselves must be regulated (Mortimore, 

et al., 1987; Mortimore and Poso, 1987; Olson, et al., 1990). Any mechanism that can 

be triggered by environmental stress lends itself to experimental manipulation. Such a 

mechanism can be modulated simply by changing experimental conditions. Furthermore, 

mechanisms required for a stress-induced degradation can be shown to be specific by 

lack of effect on basal mechanisms. For example, cells respond to starvation by 

increasing the degradation of long-lived proteins. 3-Methyladenine is a drug that 

specifically blocks the enhanced degradation without affecting basal degradation. 3- 

Methyladenine is a potent inhibitor of autophagy, a mechanism for delivering proteins to 

lysosomes for degradation. Such results provide evidence that autophagy plays a role in 

enhanced degradation but not basal degradation of proteins. 

Background for Fructose 1.6-Diphosphate Aldolase B 

In the next section, I discuss potential mechanisms for the degradation of the 

liver isoform of fructose 1,6-diphosphate (FDP) aldolase, called aldolase B. Molecular 

mechanisms for the degradation of aldolase B have not been examined, but preliminary 

examination indicated potential roles for ubiquitination and signal-mediated delivery to 

lysosomes. Before examining the degradation of aldolase B, this section serves to 

familiarize the reader with aldolase B in relation to other aldolase isoforms. There are 

two major classes (I and II) of FDP aldolase that have no sequence homology 

(Alefounder et al., 1989) and utilize different catalytic mechanisms (Rutter et al., 1966). 

Class II aldolases are only found in micoorganisms and are considered no further here, 

but class I aldolases are represented in all taxons (Rutter et al., 1966). Microorganisms 

variably lack class I aldolase. For example, some strains ofE.coli contain class I 

aldolase (Alefounder et al., 1989), whereas strain JM83 lacks class I activity (Sakakibara 

et al., 1989). Protozoa and muticellular eukaryotes contain class I aldolases. 

The FDP aldolase isoforms of mammals are the best characterized. Most studies 
of FDP aldolase degradation examine mammalian isoforms. Mammalian aldolase 
isoforms are synthesized from three separately regulated genes coding different proteins. 
Muscle, liver, and brain each express one predominant isoform designated A, B, and C, 
respectively (Rutter et al., 1966). As such, liver aldolase and aldolase B are synonyms. 
Likewise, aldolase A is also called muscle aldolase, and aldolase C is referred to as brain 
aldolase. Most tissues, including embryonic, contain combinations of aldolase A and C, 
but aldolase B appears to be exclusively expressed in liver and kidney cells (Rutter et al., 

Aldolase classes and isoforms are distinguished by clear differences in their 
enzymatic properties. For example, class I aldolase like that in mammals is totally 
functional in the presence of EDTA, but class II enzymes of E. coli and other microbes 
are completely inhibited. The three mammalian isoforms (A, B, and C) can be 

distinquished by differences in specific activity, sensitivity to carboxypeptidase A, and 

kinetics (V^and K m ) for different substrates (Rutter et al., 1966). 

Aldolases A, B, and C also are characterized by distinct native epitopes. Thus, 
antibody to one poorly recognizes the others. However, an antibody against a specific 
isoform for one animal will similarly recognize the same isoform of a different species, at 
least amongst mammals (Penhoet and Rutter, 1975). Native and denatured epitopes of 
aldolase B have also been shown to be distinct from each other (Reznick et al., 1985). 
Chemical denaturation of aldolase B before immunization resulted in antibody that failed 
to immunoprecipitate native enzyme activity but could precipitate degradative fragments 
of aldolase B. Since aldolase has stable structure that spontaneously refolds into native 
conformations, only fragments sufficiently denatured by degradation were detected by 
antibody against the denatured aldolase B (Reznick et al., 1985). Anti-native aldolase B 
had converse immunoreactivity. Thus, it was proposed that three-dimensional 
conformation is important for antibody recognition of native surface epitopes, whereas 
the native structure buries and masks denatured epitopes (Reznick et al., 1985). 

The three dimensional structure of all class I aldolase isoforms is conserved from 
bacteria to humans (Alefounder, et al., 1989). Secondary and tertiary structures of 
aldolase are very stable. This is true at the quaternary level, too. Aldolase occurs as 
very stable tetramers that do not undergo subunit exchange after synthesis (Lebherz, 
1975; Lebherz, 1972). Different aldolase isoforms co-synthesized in the same cell 
randomly associate into stable heterotetramers. Thus, immunoprecipitation with 

antibody against one isoform specifically precipitates antigenically unrelated isoforms in 

the same tetramer. However, surface charge and pi on different isoforms varies. Thus, 

isozymes containing different ratios of two isoforms (e.g. A4, A 3 B, A2B2, AB 3 , or B 4 ) 

can be separated by isoelectric focusing (Lebherz, 1972; Penhoet and Rutter, 1975). 

Comparison of X-ray crystallographic results shows that secondary and tertiary structure 

between muscle, liver, and Drosophila aldolases are very close (Berthiaume et al., 1993). 

Aldolase isoforms vary in their capacity to bind actin cytoskeletons. In the only 
paper to measure cytoskeletal association of all three isoforms (A B, and C), 
investigators claimed that different purified isoforms of aldolase had tissue-specific 
affinity for cytoskeletal preparations isolated from different tissues (Kusakabe et al., 
1997). After mixing crude cytoskeletons with a known amount of purified aldolase, they 
pelleted the mixture and then only measured unbound activities in supernatants. They 
failed to show removal of endogenous aldolase from cytoskeletal preparations, so 
measurements could be contaminated and include competitive effects. However, their 
results were consistent with other investigators in that relative tightness of binding to 
actin cytoskeleton is greatest to least, aldolase A aldolase B, then aldolase C (Clarke, et 
al., 1982; O'Reilly and Clarke, 1993). 

Four "isotype specific" sequences contain most of the variation and the carboxyl 
terminus has the greatest diversity (Marchand et al., 1988; Paolella et al., 1984; 
Rottmann et al., 1984). The carboxyl terminus is important in determining isoform- 
specific catalytic properties (Berthiaume et al., 1993; Gamblin et al., 1991; Penhoet and 

Rutter, 1975). Aldolase B has the lowest specific activity amongst aldolase isoforms and 

is least sensitive to proteolytic alterations in this region. It can lose up to four C-terminal 

residues without affect its enzymatic activity (Berthiaume et al., 1993; Horecker et al., 

1985). Even with 10 to 20 residues removed by carboxypeptidase, aldolase B retains 

almost half its activity (Penhoet and Rutter, 1975). However, aldolase A absolutely 

requires a carboxyl terminal tyrosine at residue 364 (Y364) for activity that is about 20 

times greater than aldolase B, and when aldolase A loses its C-terminus the remaining 

activity resembles that of aldolase B (Takahashi et al., 1989; Gamblin et al., 1991). 

These results indicate that alterations in the carboxyl terminus of aldolase B (such as 

epitope tagging) are less likely to affect its properties than other aldolase isoforms. 

All fructose 1,6-diphosphate aldolase enzymes catalyze a reversible reaction 

essential for glycolysis and gluconeogenesis. Aldolase B is the liver form of this enzyme 

expressed to the exclusion of other forms of aldolase in normal hepatocytes (Asaka, et 

al., 1983). Since liver is the only organ known to export glucose (Stein and Arias, 1976; 

Stryer, 1988), aldolase B performs gluconeogenesis for the entire body. Liver also 

provides amino acids during starvation and in three days can lose nearly half its weight 

(and protein content), a faster loss than other tissues (Wing et al., 1991). In this regard, 

aldolase B is an example of an abundant cytosolic protein that undergoes enhanced 

degradation during starvation which yields amino acids for export to other organs. Liver 

amino acids can also be converted to glucose or ketone bodies to provide energy sources 

during starvation. Abundant long-lived liver enzymes that mediate 


glycolysis/gluconeogenesis, like aldolase B and glyceraldehyde phosphate dehydrogenase 
(GAPDH) are poised between two mutually exclusive functions: catalyzing carbohydrate 
metabolism and providing amino acids for protein biosynthesis or energy metabolism. 

Liver and kidney are the only tissues having predominant aldolase B expression 
(Penhoet and Rutter, 1975). Both organs demonstrate enhanced degradation of proteins 
during amino acid starvation (Olsen, et al. 1990). Liver and kidney also receive the 
largest fraction of the body's basal blood flow, 27% and 22%, respectively, followed by 
15% for muscle and 14% for brain (Guyton, 1979). This is consistent with involvement 
of the two former organs in regulating serum components and contribution of aldolase B 
to serum glucose and amino acids during starvation. Among aldolase isoforms, aldolase 
B contributes a greater role in carbohydrate and protein metabolism that is not limited to 
local cells and tissues, but extends to the entire body. 

Mechanisms for Degradation Aldolase B 
Inactivation by Limited Proteolysis 

Alteration of aldolase A and B's carboxyl termini was proposed to down- regulate 
activity (Pontremoli et al., 1982; Pontremoli et al., 1979). During starvation, aldolase B 
activity is lost from liver faster than loss of immunoreactivity. Thus, investigators 
suggested that starvation-induced inactivations precede total degradation of aldolase A 
and B, providing more rapid down-regulation of activity (Pontremoli et al., 1979). 
Inactivation happens by limited C-terminal cleavage that can maintain native 

immunoreactivity and barely affect mobility on SDS-PAGE. One group of investigators 

proposed phosphorylation near the C-terminus of aldolase as an inactivating mechanism, 

but this was only demonstrated in vitro (Sygusch et al., 1990). More likely inactivation 

occurs by limited proteolysis which would have a much more profound impact on 

aldolase A activity than on aldolase B activity (previous section, last paragraph). 

The best characterized mechanism for aldolase inactivation is limited proteolysis 
by a dipeptidyl (two residues per cleavage) carboxypeptidase on lysosomes (Pontremoli 
and Melloni, 1986; Horecker et al., 1985). The peptidase, cathepsin M, was defined as 
a cathepsin B or L-like activity associated with the cytosolic surface of lysosomal 
membranes. During starvation, a lysosomal matrix cathepsin B/L associates with 
lysosomal membranes, acquires activity at neutral pH, and becomes exposed to the 
cytosolic compartment as cathepsin M (Pontremoli et al., 1984; Pontremoli et al., 1982). 
Specific cleavage sites have been characterized in vitro (Horecker et al., 1985). 

Starvation-induced in vivo loss of liver aldolase specific activity correlated with 
loss of carboxyl terminal tyrosine residues which was estimated by isolating aldolase B 
and measuring lost tyrosine content in an acid soluble peptide released from the C- 
terminus with subtilisin (Pontremoli et al., 1982). According to such experiments, 
inactivated aldolase B constitutes about 40% of the aldolase in liver after 60 hours of 
starvation. Most of the inactivated aldolase B must occur in cytosol, because only a 
small fraction of total aldolase (about 10%) is associated with pelletable fractions from 
liver (Kominami et al., 1983; Kopitz et al., 1990). Moreover, intralysosomal 

degradation of aldolase is rapid (see below), precluding accumulation of an inactivated 

form in such organelles. The results are consistent with inactivation occurring in the 

cytosolic compartment, albeit by an activity associated with the cytosolic surface of 


In Vitro Denaturation of Aldolase and Need for In Vivo Mechanism 

Except for 20 "loose" amino acid residues at the carboxyl terminus, the stability 

of aldolase structure resists proteolysis and requires denaturation for rapid in vitro 

proteolysis to proceed. In optimized conditions with cathepsin D, only about 20 amino 

acids of aldolase A can be digested from its carboxyl terminus (Offermann et al., 1983). 

In vitro proteolysis with either meprin (a metalloproteinase) or a mixture of lysosomal 

proteases produces only a slight increase in SDS-PAGE mobility, and the remaining part 

of aldolase A has a thermal stability identical to the native enzyme (Bond and Offermann, 

1981). Purified aldolase B digested with a lysosomal extract also only undergoes 

limited proteolysis, losing some but not all its activity (Chappel et al., 1978). However, 

denaturing pretreatment with disulfides like glutathione (Offermann et al., 1983) or 

cystine (Bond and Offermann, 1981) permits extensive proteolysis to occur. Given 

this, there must be a "denaturing" mechanism in vivo to allow degradative turnover of 

aldolase to occur. Interestingly, aldolase B sequestered in vivo and isolated with 

lysosomes is susceptible to more extensive in vitro proteolysis in the lysosomal 

preparations (Kominami, et al., 1983; Ueno and Kominami, 1991). Apparently, 


aldolase B becomes sensitized to proteolysis by a mechanism in cytosol before 
sequestration or in intact lysosomes after sequestration. 

After loading aldolase A into endosomes at 19°C, temperature can be raised to 
37°C allowing rapid fusion of endosomes with lysosomes. Thus, intralysosomal 
degradation can be measured. By this method, native or variously denatured and 
inactivated aldolases all degrade rapidly with similar rates (ty 2 < 10 min). Since its Xy 1 is 
normally many hours in cytosol, sequestration appears rate limiting for lysosomal 
degradation of aldolase (Bond and Aronson, 1983). The results of the endocytic loading 
experiments indicate that a mechanism for denaturing and sensitizing aldolase to 
proteolytic attack can occur in lysosomes or other organelles of the endosomal pathway. 
Thus, a cytosolic denaturing mechanism is not necessary for intralysosomal degradation 
of aldolase, but a role in delivery of aldolase to lysosomes cannot be excluded. 

The tetrameric structure of aldolase is well established (Lebherz, 1972). This 
quaterenary structure seems important for aldolase stability. Recently, Beernink and 
Tolan have indentified specific amino acids that mediate subunit interaction between 
aldolase monomers (Beernink and Tolan, 1996). Significantly, a mutant with only two 
amino acid changes retains enzymatic activity but exists as monomers. These monomers 
(and dimers created with single amino acid mutations) are more sensitive to chemical or 
thermal inactivation, indicating "looser" structure. Thus, tetrameric association 
improves structural stability. 

Lysosomes are acidic inside (pH ~5), and reversible in vitro dissociation of 

aldolase into monomers occurs at pH < 6.0 (Beernink and Tolan, 1996). Acidic pH 

affecting adolase structure is also indicated by reduced enzymatic activity. Thus, 

intralysosomal pH would have a denaturing effect that could permit lysosomal 

proteolysis. However, other investigators incubated aldolase B with crude lysosomal 

hydrolases at acidic pH and failed to get significant proteolysis (Chappel et al, 1978). 

Apparently, low pH is insufficient to permit further proteolytic attack, and aldolase 

denaturation must require other factors. Consistent with this, lysosomes purified from 

liver contain detectable aldolase B which is susceptible to proteolysis when the intact 

lysosomes are incubated in vitro at pH < 5.5 (Kominami, et al., 1983; Ueno and 

Kominami, 1991). The endocytic loading experiments described above indicate 

lysosomes (or an endocytic compartment) must contain denaturing factors, but this does 

not exclude the possibility of a cytosolic denaturation of aldolase B before delivery to 


Aldolase A has been radiolabeled, inactivated and denatured, then microinjected 

into cultured cells (Hopgood et al., 1988; Knowles et al., 1989). The procedure delivers 

the enzyme into cytosol where it normally resides. As with endocytic loading, 

degradation rates for aldolase were similar whether the enzyme was native, inactivated, 

or denatured. Denaturation of aldolase is not rate limiting for degradative steps before 

lysosomes as well as within them. Degradation of aldolase microinjected into cytosol 

matched expected turnover for aldolase (t Vl = 30 hours) which was much slower than for 

aldolase loaded into lysosomes (t./ 2 < 10 minutes). Assuming that degradation occurs 

within lysosomes, this suggest that sequestration of aldolase is rate-limiting for its 

turnover (Bond and Aronson, 1983; Bond and Offermann, 1981; Hopgood et al., 1988; 

Knowles et al., 1989). 

Though in vitro studies indicate denaturation of aldolase structure is necessary 
for proteolysis, in vivo denaturation is not rate-limiting for delivery to or degradation 
within lysosomes. These data support a model in which aldolase delivery to lysosomes is 
rate limiting followed by rapid intralysosomal proteolysis which would need a faster 
denaturing mechanism. Lysosomal acidity might facilitate denaturation of aldolase, but 
acidity alone is insufficient for sensitizing stable aldolase structure to attack by acid 
hydrolases. The above data do not exclude a cytosolic denaturing mechanism for 
aldolase, but indicate that such a mechanism is not rate limiting and not necessary for 
intralysosomal proteolysis. The next two sections review mechanisms for the delivery of 
cytosolic proteins to the lysosomal lumen, a process that appears rate-limiting for 
aldolase degradation. 

Autophagy is the sequestration of cytoplasm into vesicles for intralysosomal 
degradation and is the only mechanism proposed for the complete degradative turnover 
of aldolase. There are two forms of autophagy: macroautophagy and microautophagy. 
Commonly, investigators use the term "autophagy" to mean macroautophagy which is 

the better characterized form. Likewise, "autophagy" used here refers to 

macroautophagy, and reference to "microautophagy" will be explicit. 

Autophagy (macroautophagy) begins with a ribosome-free portion of 
endoplasmic reticulum engulfing a portion of cytoplasm. Autophagy non-selectively 
sequesters cytosol and organelles into distinct autophagic vacuoles. The autophagic 
vacuoles mature including a process of acidification. Finally, mature autophagic 
vacuoles fuse with lysosomes producing autolysosomes in which degradation occurs 
(Dunn, 1990; Dunn, 1990). Enhanced autophagy is initiated by amino acid starvation 
and is also regulated by hormones (Hendil et al., 1990; Seglen and Bohley, 1992). In the 
model of Figure 1-1, non-selective autophagy is represented by the upper pathway in the 
diagram. The lower pathway of Figure 1-1 (Receptor-Mediated Targeting) is discussed 
in the next section. 

Microautophagy seems simpler than macroautophagy. During microautophagy 
the lysosomal membrane itself invaginates, extending a finger of cytosol into the 
lysosome. This protrusion pinches off producing an intralysosomal vesicle that gets 
degraded with its cytosolic content. Apparently, microautophagy can occur in vitro, but 
the complexity of macroautophagy has not been reconstitiuted (Seglen and Bohley, 
1992). Unlike macroautophagy which has discrete autophagic vacuoles, microautophagy 
fails to produce separable organellar compartments. Thus, microautophagy requires 
time-consuming electron microscopy to demonstrate its exsistence and remains poorly 
characterized (Seglen and Bohley, 1992). 


Figure 1-1: Mechanisms for Stress-Induced Degradation of Cytosolic Proteins in 
Lysosomes . Autophagy (upper pathway) and receptor-mediated targeting (lower 
pathway) were proposed for stress-induced delivery of cytosolic proteins to lysosomes 
for degradation; the arbitrary cytosolic protein is shown as a tetramer (aldolase B occurs 
as a tetramer); components of the pathways are labeled on the diagram; processes are 
labeled by boxed numbers: 1, association with or engulfment by autophagic membranes; 
2, sequestration into double-membrane bound autophagic vacuole; 3, maturation of 
autophagic vacuole (acidification and acquisition of lysosomal hydrolases); 4, proteolysis 
into polypeptide fragments; 5, complete degradation to amino acids; 6, disassembly and 
denaturation of structure by an unknown factor; 7, association with a receptor complex 
on the lysosomal surface; 8, translocation across the lysosomal membrane. 

Selective mechanisms of autophagy exist. Methylotrophic yeast use a selective 
mechanism of autophagy to degrade peroxisomes when switched from methanol to a 
different carbon source, and electron microscopic morphology shows a mechanism 

topologically identical to microautophagy (Tuttle et al., 1993). Occurrence of selective 

microautophagy in higher organisms has not been demonstrated, and a role for 

microautophagy in degradation of aldolase has not been studied. If selective autophagy 

does occur for aldolase, then a receptor-mediated complex would be required for 

selectivity. Such a receptor complex could form on the lysosomal surface (Fig. 1-1, 

lower pathway), followed by microautophagic sequestration. However, receptor 

function does not distinguish microautophagy and macroautophagy, so Figure 1-1 only 

distinguishes non-selective autophagy (upper pathway) from a hypothetical selective 

process that might include microautophagy (lower pathway). 

Autophagy (macroautophagy) is a subject of active research, producing almost 

300 related papers in just the last five years. FDP aldolases are generally abundant, 

commonly known, cytosolic enzyme, and the muscle isoform, aldolase A, is 

commercially available. Aldolase A and aldolase B have been used as markers for 

autophagic uptake of cytosol into lysosomes, and degradation of aldolase by autophagy 

is well established (Henell et al., 1987; Kominami et al., 1983; Kopitz et al., 1990; 

Seglen and Gordon, 1982; Ueno et al., 1990). Per O. Seglen's laboratory briefly treated 

starved hepatocytes with cycloheximide to prevent new protein synthesis and estimated 

degradation rates by loss of enzyme activity. Incubations were short to avoid depletion 

of autophagic factors (continual autophagy requires new protein synthesis). A lysosomal 

inhibitor (leupeptin) was used to estimate how much degradation occurred in lysosomes. 

In this way, starvation-induced degradation of aldolase B was lysosomal occurring at 

3.6±0. 1 %/h. Other cytosolic en2ymes with widely different half-lives were similarly 

tested. As expected, they had very different total degradation rates. However, 

lysosomal degradation (3.3-5.3 %/h) and rates of accumulation in organelles during 

lysosomal inhibition (3.1-3.7 %/h) were similar for all the enzymes. These rates match 

rates of starvation-induced autophagy (3-5 %/h) and were blocked with 3-methyladenine 

a specific inhibitor of autophagy. Thus, Seglen concluded that cytosolic enzymes, 

including aldolase B, are degraded via non-specific autophagy (Kopitz et al., 1990). 

Interestingly, of all the enzymes tested by Seglen, only two were exclusively degraded in 

lysosomes, aldolase B and lactate dehydrogenase H (Kopitz et al., 1990). 

Coincidentally, both aldolase B and lactate dehydrogenase H contain sequence motifs for 

receptor-mediated targeting to lysosomes for degradation. 

Receptor Mediated Targeting to Lysosomes 

A pentapeptide sequence (KFERQ) of RNAse A was shown to mediate its 

delivery to lysosomes for degradation during nutrient deprivation (Dice and Chiang, 

1988). Characterization of this signal identified a motif contained in a subset of cellular 

proteins that undergo enhanced degradation in lysosomes during nutrient deprivation 

(Wing et al., 1991). The motif has been proposed as a binding site for a molecular 

chaperone called HSC73 which then delivers motif-containing proteins into lysosomes in 

an ATP-dependent manner (Chiang et al., 1989). The mechanism (Figure 1-1, lower 

pathway) also requires intralysosomal HSP70, and a recently identified lysosomal 

membrane receptor, LGP96 (Cuervo, et al. 1996). J. Fred Dice proposed that this 


pathway occurs by a mechanism analogous to transmembrane transport of proteins 

during organellar biogenesis (Dice and Chiang, 1988; Terlecky et al., 1992; Wing et al., 

Recently, Dice's group made a major advance by identifying a receptor protein in 
the target membrane of lysosomes that mediated transmembrane translocation (Cuervo 
and Dice, 1996). The lysosomal membrane protein, LGP96, was demonstrated to be a 
rate limiting component in this degradative pathway. CHO cells overexpressing human 
LGP96 by two to three fold had correspondingly increased degradation of long-lived 
proteins. Furthermore, lysosomes isolated from these cells had two to three fold higher 
ATP-dependent uptake of the glycolytic enzyme GAPDH (glyceraldehyde 3 -phosphate 
dehydrogenase), a known substrate for his pathway (Cuervo and Dice, 1996) . 
Unfortunately, GAPDH does not contain sequence matching previously established 
criteria for the receptor-mediated targeting motif. The previous criteria define necessity 
for an "essential" glutamine, but in GAPDH, asparagine apparently can substitute for 
glutamine (Dice, personal communication). As just described, receptor function has been 
demonstrated in living cells; however, evidence for in vivo function of the signal peptide 
is lacking. 

A recent study found that the conformation of signal motifs were inappropriate to 
mediate the receptor-mediated lysosomal targeting pathway (Gorinsky et al., 1996). 

Some proteins known to contain motifs for the pathway, including RNAse A, also have 
known three dimensional structures. The peptide signal motifs are supposed to be 

recognized by cytosolic HSC73 which is required for delivery to lysosomes. However, 

signals on proteins of known structure are either embedded or occurred in a-helical 

conformations. Since hsp70-type chaperones require extended conformations for 

recognition, the investigators concluded that HSC73 would required other unknown 

factors to relax a substrate protein's structure and allow signal-mediated targeting to 

lysosomes to occur (Gorinsky et al., 1996). The lower pathway in Figure 1- 

1 summarizes receptor-mediated targeting to lysosomes, including an unknown factor 

that alters the structure of the cytosolic protein. 

Except for the work presented in this dissertation, no studies have examined any 
aldolase isoform as a substrate for the receptor-mediated pathway. Aldolase A binds 
very well to GAPDH presumably for greater glycolytic efficiency (Verlessy and Vas, 
1992). Both these proteins appear to be regulated at similar high concentrations in cells 
(Verlessy and Vas, 1992), and aldolase B and GAPDH undergo similar starvation- 
induced degradation in cultured cells (Kopitz, et al., 1990). Does aldolase B follow 
receptor-mediated targeting to lysosomes like GAPDH? 

All vertebrate aldolases contain a conserved motif (Fig. 1-2, residues 12-16) for 
receptor-mediated targeting of cytosolic proteins into lysosomes (Dice and Chiang, 
1988; Zhang et al., 1995). In mammalian aldolase B, two additional sequences for the 
lysosomal targeting motif were found (Fig. 1-2, residues 58-62 and 107-1 1 1). Whether 
any of these three motifs are functional was unknown. Though the lysosomal targeting 
motif has rather broad criteria (Dice and Chiang, 1988), an aldolase-sized (40 kD) 



1 Q58 



44 QUI 















354 364 

Figure 1-2: Amino Acid Sequences of Aldolase B Isoforms Used in This Study . The 
upper and lower sequences are for rat and human liver aldolase, respectively. Boldface 
indicates non-identical residues. Underline indicates pentapeptide signal motifs for 
receptor-mediated lysosomal targeting. Large arrows point to essential glutamines of the 
signals as indicated (sequence data from Tsutsumi et al., 1984 and Paolella et al., 1984). 

protein would have only a 7% chance of randomly containing three such signal motifs. 

In addition to the motifs, aldolase B has properties similar to other substrates of this 

signal-mediated degradative mechanism: (1) long-lived, (2) cytosolic, (3) housekeeping 

protein, and (4) degraded in lysosomes by enhanced proteolysis during nutrient 

withdrawal. Furthermore, aldolase is closely associated with another glycolytic enzyme 

glyceraldehyde-3-phosphate dehydrogenase (GAPDH) which is an established substrate 

for receptor-mediated targeting to lysosomes (Aniento et al., 1993). Aldolase and 

GAPDH form a complex that facilitates their sequential roles in glycolysis (Verlessy and 

Vas, 1992), both are very abundant, and their in vivo turnover rates are very similar 

(Kuehl and Sumsion, 1970), suggesting that they could share degradative mechanisms. 

Furthermore, the receptor-mediated pathway proceeds by transmembrane translocation 

into lysosomes by a mechanism like that of organellar biogenesis. Coincidentally, the 

aldolase isoform of Trypanosoma brucei (45% identical to aldolase B) undergoes 

transmembrane transport during biogenesis of the unique glycolytic organelle 

of this protozoan (Marchand et al., 1988). Together, these facts suggested that aldolase 

B was a likely candidate for receptor-mediated targeting to lysosomes via the proposed 

transmembrane transport mechanism. 

Ubiquitination and the Degradation of Long-lived Proteins 

Ubiquitination is an orderly process whereby a 76-amino acid polypeptide, 

ubiquitin, is covalently conjugated to other proteins at its carboxyl terminus. In a series 

of transfers, three enzymes (El, E2, and E3) covalently bind and pass ubiquitin to the 

next protein. El, called ubiquitin-activating enzyme, first conjugates ubiquitin's carboxyl 

terminus. This step is obligatory, and cell lines with temperature-sensitive ubiquitination 

have defects traced to mutations in El (Kulka, et al., 1988; Chowdary, et al., 1994). El 

transfers ubiquitin to an E2 which transfers it to an E3 which finally conjugates the 

ubiquitin to a target protein (sometimes, the E3 step is skipped). Most protein 

ubiquitination requires a single El protein, but E2 and E3 enzymes occur as families that 

regulate and confer specificity for ubiquitination. This arrangement explains why genetic 

defects in general ubiquitination only occur in El enzymes (Ciechanover and Schwartz, 

1994; Hochstrasser, 1992). 

Cells die without ubiquitination. The process has been implicated in a wide 
variety of cell functions reviewed elsewhere (Hochstrasser, 1996; Jentsch, 1992). 
Ubiquitination was originally discovered in the rapid degradation of short-lived and 
abnormal proteins by cytosolic proteases, and this role remains the best characterized 
(Hershko and Ciechanover, 1992). Heat stress causes enhanced degradation of long- 
lived proteins in E36 Chinese hamster lung cells. This heat-stress induced degradation 
occurs in lysosomes via autophagy and requires ubiquitin-activating enzyme El (Gropper 
et al., 1991; Handley-Gearhart et al., 1994). However, specific long-lived proteins that 
utilize this ubiquitin-mediated autophagic mechanism have not been identified. 

Whether for short-lived or long-lived proteins, ubiquitin-mediated turnover 
involves attachment of multiple ubiquitins on a protein targeted for degradation 
(Hershko and Ciechanover, 1992). A ubiquitinated protein (a.k.a. ubiquitin conjugate) is 

a substrate for further ubiquitination, and additional ubiquitins preferentially conjugate to 

already attached ubiquitin. A chain of ubiquitins is built on a protein to be targeted. The 

multiubiquitin chain then acts as a signal for degradation of the targeted protein. Each 

ubiquitin adds an additional 76 amino acids to the protein, and successive intermediates 

of multiubiquitination can be demonstrated as a ladder of bands on SDS-PAGE that 

contain both ubiquitin and the targeted protein (Chau et al., 1989; Hershko and 

Ciechanover, 1992). Multiubiquitination is a well-established signal for stress-induced 

degradation of short-lived proteins by a major protease complex in cytosol. Though 

total long-lived proteins undergo ubiquitin-mediated stress-induced degradation in 

lysosomes, a role for multiubiquitination has not been established for lysosomal 

degradation of any specific cytosolic protein. 

Recently, work in the laboratory of William A. Dunn, Jr. demonstrated a 
connection between ubiquitin and the long-lived protein aldolase B. The evidence 
includes data presented in this dissertation, two meeting abstracts, and a manuscript 
which has been submitted (Lenk et al., Submitted 1998; Susan and Dunn, 1996; Susan et 
al., 1995). Together the data support that aldolase B is multiubiquitinated in vivo and 
suggest that ubiquitination is involved with stress-induced autophagic degradation of 
aldolase B in lysosomes. 

S. E. Lenk and William A. Dunn, Jr. provided the first evidence that aldolase B 
has ubiquitinated forms (Figs. 1-3 and 1-4), including a major 68 kD form (Ub68) 
enriched in lysosomes during nutrient deprivation (Lenk et al., Submitted 1998; Susan et 

Figure 1-3: Characterization of A Major Ubiquitin-Protein Conjugate Enriched in 
Autophagic Vacuoles , a) Rats were starved to induce autophagy and lysosomal uptake 
of ubiquitinated proteins. Subcellular fractions of liver were prepared, and equal protein 
from cytosolic (Cy), lysosome-enriched (Ly), and autophagic vacuole-enriched (AV) 
fractions were run on SDS-PAGE, western blotted, and labeled with antibody against a 
major ubiquitinated protein (anti-Ub68). Note major bands at 68 kD and cross-reactivity 
to a 40 kD protein in cytosol. The 40 kD protein was identified as aldolase B by peptide 
sequence analysis, b) Cytosol was circulated on an anti-Ub68 column, washed, eluted, 
and preparative SDS-PAGE performed. A gel strip was stained with Coomassie blue R- 
250 (CB), and the remaining gel was blotted and cut into strips individually stained with 
anti-Ub68 (Ub68) or anti-ubiquitin (Ub). Ub68 and aldolase are indicated at 68 kD and 
40 kD, respectively. Arrowheads indicate positions of bands suggestive of intermediates 
of multiubiquitination of aldolase (from Lenk, et al., Submitted 1998). 

a. Cy Ly AV 


68 ► 


40 ► 

b. CB 





Ub68 Ub 





Figure 1-4: Amino Acid Starvation Increases Lysosomal Association of Putative 
Ubiquitinated Aldolase B via Autophagy . Fao rat hepatoma cells were incubated on 
media with or without amino acids (AA) and the autophagy inhibitor 3-methyladenine 
(3MA) as indicated. Sub-cellular fractions were collected, and equal protein of 
lysosome-enriched fractions were run on SDS-PAGE, western blotted, and stained with 
antibodies against Ub68 and ubiquitin (Ub). Positions of molecular weight standards and 
Ub68 (arrowhead) are indicated (from Lenk et al., Submitted 1998). 




205 — 

116 — 

45 — 



al., 1995). Aldolase A and B are long-lived proteins known to undergo degradation by 

autophagy (Henell et al., 1987; Kominami et al., 1983; Kuehl and Sumsion, 1970; Ueno 

and Kominami, 1991; Ueno et al., 1990). It has been determined that amino acid 

deprivation (starvation) rapidly enhances autophagy in cultured cells (Kopitz et al., 1990; 

Seglen and Gordon, 1982; Ueno et al., 1990). Since Ub68 increases in lysosomes under 

similar conditions (Fig. 1-4), the evidence suggested that ubiquitination might play a role 

in stress-induced autophagic degradation of aldolase B. 

Hypothesis for Stress-Induced Degradation of Aldolase B 

The field of protein degradation has made great progress in determining 
molecular mechanisms for the degradation of short-lived proteins via a cytosolic protease 
complex (the proteasome); however, long-lived proteins which are generally thought to 
be degraded in lysosomes have relatively poorly characterized degradative mechanisms. 
Cellular degradative mechanisms that respond to environmental changes facilitate 
experimental characterization. Starvation (amino acid and serum deprivation) and heat 
stress can induce regulated mechanisms for the degradation of long-lived proteins. 
Figure 1-1 presents two pathways proposed for the stress-induced delivery of cytosolic 
proteins to lysosomes for degradation: autophagy or receptor-mediated targeting (lower 

In the simplified diagram of Figure 1-1, only topological changes in autophagy 
(upper pathway) are shown for a tetrameric cytosolic protein sequestered into an 
autophagic vacuole (steps 1 and 2) that fuses with lysosomes (step 3). Heat stress 

induces autophagic degradation that requires ubiquitination, but specific proteins that are 

ubiquitinated during stress-induced autophagy have not been identified. Aldolase B is 
known to undergo autophagy, and a putative ubiquitinated form aldolase B associates 
with autophagic vacuoles and lysosomes during starvation. To establish a specific 
protein for ubiquitin-mediated autophagy, we examined aldolase B as a likely substrate 
for ubiquitin-mediated autophagy. 

In Figure 1-1, receptor-mediated targeting to lysosomes is also drawn showing 
required components (lower pathway). Since established substrates for this pathway 
have conformations that would prevent receptor recognition, unknown factors (smallest 
circles) have been proposed to relax the structure of substrate proteins (first arrow) 
which probably includes disassembly of subunits from quaternary structures (rough- 
drawn oval with small circles attached). An exposed signal then mediates assembly of a 
complex on the lysosomal surface (second arrow), including the HSC73 chaperone 
(medium gray square), the substrate protein (extended coils), the lysosomal membrane 
protein LGP96 (darkest rectangle), and possibly other factors (small circles). 
Transmembrane translocation (third arrow) also requires an intralysosomal HSP70 
chaperone (dark gray square). Aldolase B has characteristics similar to known substrates 
for receptor-mediated targeting to lysosomes, but this mechanism was not examined for 
any aldolase. Evidence will be shown that ubiquitinated forms of aldolase B have a more 
denatured conformation. If aldolase B follows receptor-mediated degradation, then 
ubiquitin could represent the unknown factor needed to relax substrate structure. 

The relationship between receptor-mediated targeting of cytosolic proteins to 

lysosomes and ubiquitin-mediated autophagic degradation had not been examined. Since 

aldolase B was a potential substrate for both pathways, I hypothesized that during stress, 

aldolase B requires both ubiquitination and a receptor-mediated targeting signal for 

enhanced degradation in lysosomes. 

General Strategy 

I adopted the hypothesis that during stress, aldolase B requires both 
ubiquitination and a receptor-mediated targeting signal for enhanced degradation. The 
two requirements in this hypothesis were separately tested: ubiquitination and a receptor- 
mediated targeting signal. In this regard, there were two corresponding aims of this 
investigation: Aim #1, perturb ubiquitination and examine the effects on stress-induced 
delivery of aldolase B to lysosomes; Aim #2, mutate potential lysosomal targeting signals 
and examine effects on starvation-induced degradation of aldolase B. 

My first aim was to determine whether stress-induced degradation of aldolase B 
requires ubiquitination. Antibodies were raised against aldolase B expressed in and 
isolated from E. coli. Since bacteria lack ubiquitin, the antibodies were produced against 
antigen that did not contain ubiquitin or ubiquitin-conjugated proteins. With the 
antibodies, the presence of aldolase B epitopes in a major 68 kD ubiquitinated protein 
(Figs. 1-3 and 1-4, Ub68) and other ubiquitin conjugates was confirmed in subcellular 
fractions from rat liver. Epitope-tagged aldolase B was expressed in E36 (parent) and 
ts20 (ubiquitination mutant) cells previously used to establish ubiquitin-dependency for 

heat stress-induced autophagic degradation of long-lived proteins. By examining 

changes in the endogenous aldolase A and exogenous aldolase B associated with 

pelletable subcellular fractions, evidence was found that these aldolase isoforms require 

ubiquitination for autophagic degradation in lysosomes during heat stress. This 

supported my hypothesis that stress-induced degradation of aldolase B requires 


An attempt was made to use protein degradation measurements to confirm that 
heat stress-induced degradation of aldolase B requires ubiquitination. Degradation of 
aldolase B was found to utilize a temperature-dependent cytosolic proteolytic 
mechanism. The cytosolic proteolysis of aldolase B at heat stress temperatures was 
similar in magnitude to induced ubiquitin-mediated autophagic degradation. Since the 
cytosolic mechanism turned out to be ubiquitin-independent, degradation measurements 
could not confirm ubiquitin-mediated degradation of aldolase B via lysosomes. 
However, the results demonstrate that mechanisms for degradation of aldolase B include 
a novel cytosolic proteolysis. 

My second aim was to test whether a receptor-mediated targeting signal was 
required for stress-induced degradation of aldolase B. A sequence motif has been 
defined for targeting cytosolic proteins to lysosomes for degradation during nutrient 
deprivation, and aldolase B contains three sequences that match the motif (Fig. 1-2). 
Depriving liver-derived cell lines of serum and amino acids causes starvation-induced 
degradation of long-lived proteins including aldolase B. Vectors were constructed 
expressing epitope-tagged aldolase B and used site-directed mutagenesis to disrupt the 

putative targeting signals. Wildtype and mutant aldolase B proteins were expressed and 

assayed for starvation-induced degradation. Starvation causes enhanced autophagic 

degradation of aldolase B expressed in cultured hepatoma cells, and this enhanced 

degradation specifically required a targeting signal that includes glutamine residue #111. 

This supported my hypothesis that stress-induced degradation of aldolase B utilizes a 

receptor-mediated targeting signal. 


Cell Lines and Culturing 
General Maintenance 

Except for temperature (see following subsections), all cell lines were maintained 

similarly using standard sterile cell culturing techniques. Except where indicated, all 

supplies were obtained from Fisher Scientific, Inc. The term "standard culture 

conditions" refers to maintenance in DMEM (Sigma #D-5648), 2.2 g/1 NaHC0 3 , and 

10% FBS (Biocell #6201-00) in a 5% C0 2 atmosphere, and the standard medium for 

stably transfected cells included 0.3 mg/ml active G418 (GIBCO BRL #1 181 1-031). 

Cultures were fed every 3-4 days and passaged before complete confluency. For 

passages, cell sheets were rinsed with DPBS (Sigma #D-5652) followed by lx trypsin- 

EDTA (Sigma #T4174) in DPBS for 4-8 minutes at room temperature or 37°C as 

needed. Passages to amplify and maintain cultures for experiments were split 1 : 10 to 

1:50 (area:area), and very fast growing lines that tolerated thin splits were done down to 

1 :80. Since trypsin/EDTA diluted 1 : 10 or more with 10% FBS did not affect cell sheets 

during 30 min. at 37°C, some thin splits (at least 1:30 into medium with 10%FBS) were 

directly plated without pelleting to remove trypsin. For cultures using this short cut, 

attachment times, spreading times, growth rates, and experimental results were 

unaltered. Passages to replenish frozen stocks were always split heavily at 1:3 to 1 :6 

from freshly thawed stocks grown to near confluency. For new frozen stocks, cells 


were suspended in medium supplemented with 10% DMSO (Sigma #D-2650), incubated 

1-2 h at minus 20°C, then at minus 80°C overnight, and stored at minus 80°C for up to 

two months or transferred to liquid nitrogen for longer storage times. 

Heat Stress-Inducible E36 Cells and Ubiquitination Mutant 

Alan Schwartz kindly provided cell lines: E36 (parent), ts20 (mutant with 
temperature sensitive ubiquitin-activating enzyme El), and ts20Elc2 (mutant rescued by 
wild-type human El) Chinese hamster lung cell lines. These cells are well characterized 
for thermal control of ubiquitin-activating enzyme El activity and together have 
demonstrated that El -mediated ubiquitination is required for heat stress-induced 
degradation of long-lived proteins ((Handley-Gearhart et al., 1994); (Handley-Gearhart 
et al., 1994); (Trausch et al., 1993); (Lenk et al., 1992); (Schwartz et al., 1992); 
(Gropper et al, 1991); (Kulka et al., 1988)). 
Starvation-Inducible Cell Lines 

William A. Dunn, Jr. provided Fao (rat hepatoma) and HuH7 (human hepatoma) 
cell lines. The Fao cell line originates from a rat hepatocellular carcinoma (Reuber, 
1961), and this derivation is well documented (Deschatrette and Weiss, 1974). Fao cells 
retain a dozen liver-specific characteristics examined by Mary C. Weiss, including some 
endogenous expression of aldolase B (Deschatrette et al., 1979; Deschatrette and Weiss, 

The HuH7 cell line was isolated from a well differentiated carcinoma of a 
Japanese man and shown to secrete 16 different plasma proteins associated with liver 

function (Nakabayashi et al., 1982). Seven expected carbohydrate-metabolizing 

activities were present in HuH7 cells, but for two of these, liver-specific isoforms, 

pyruvate kinase L and a low-K m hexokinase, were not detected (Nakabayashi et al., 

1982). BHK (baby hamster kidney), and NRK (normal rat kidney) cell lines were 

examined briefly during transient transfections. 

Plasmid Vector Construction and Mutagenesis 

General Molecular Biological Methods 

Basic methods were performed essentially as described in Current Protocols in 
Molecular Biology (Ausubel, et al., 1994). Except where noted, all supplies came from 
Fisher Scientific, Inc. Kits for DNA preparations were from Qiagen and Promega. 
Restriction digestions, ligations, other DNA modifications, and PCR utilized supplies 
from Promega and New England Biolabs, except as noted below. 
PCR Primers and DNA Sequencing 

At the University of Florida, the DNA Synthesis Core Laboratory provided all 
oligonucleotide primers that William A. Dunn, Jr. or Peter P. Susan designed for PCR. 
The University of Florida DNA Sequencing Core Laboratory sequenced parts of 
plasmid vectors that were altered, or we did DNA sequencing with a Sequenase™ kit 
(U.S. Biochemical Corporation). 
Expression Vectors for Epitope-Tagged Aldolase B 

Kiichi Ishikawa provided pRAB 1710 Amp+ (Tsutsumi et al., 1984), a plasmid 
containing the cDNA of rat aldolase B (RAB) used as a PCR template in a reaction 
containing two primers (Fig. 2-1). PCR solutions were prepared according to 

manufacturer's specifications (Promega) and then run through 40 cycles (each cycle: 1 

min. at 94°C, 1 min. at 52°C, and 2 min. at 72°C), yielding DNA coding for rat aldolase 

B tagged with the 9E10 myc epitope at the carboxyl terminus (RABM). After 

restriction, the product was ligated into Xhol and Xbal sites of the vector pMAMneo- 

blue (CLONTECH), yielding pRABM which failed to express RABM. Using EcoRV 

and Xbal, I transferred the RABM code to pcDNA3 (Invitrogen), yielding 

pcDNA3RABM (Fig. 2-2). William A. Dunn, Jr. provided a pcDN A3 -based vector, 

pHAHAB which expresses human aldolase B (Sakakibara et al., 1989) with an amino 

terminal 12CA5 HA epitope, HAHAB (Lenk et al., Submitted 1998). 




b. WID9, 3' primer: TGCTCTAGACTActacaagtcttcttcagaaataagcttttgttcctcG7>lGG- 

Figure 2-1 : Primers for PCR Amplification of Insert Containing cDNA for RABM 
Expression, a) 5' primer also called, WID5; b) 3' primer also called WID9. Orientations 
are relative to 5' to 3' convention. Italics indicate reverse complementary code for 
carboxyl terminal amino acids; lower case letters indicate reverse complementary 
sequence for myc (9E10) tag; single underline indicates reverse complementary sequence 
for a stop codon; boldface indicates Xhol and Xbal restriction sites; other bases are the 
same as vector sequences. 

3.. 5' DNA Sequence: 

EcoR V Xhol start 

(aldolase B) 

D. 3' DNA Sequence: 

myc-epitope code 

C-terminal stop Xbal 

tyrosine codon 
(aldolase B) 

C. Carboxyl terminal amino acids of RABM: 


FIGURE 2-2: DNA Sequence of pcDNA3RABM a) 5' insertion site and b) 3' insertion 
site showing new sequence generated by vector construction. Boldface designates 
cDNA sequence of rat aldolase B (upper case) and human myc epitope 9E10 (lower 
case). Italics designate DNA sequence from the multicloning site of pMAMneo-blue™. 
Standard typeface designates DNA sequence in the multicloning site of pcDNA3™ 
(Invitrogen). Underline designates indicated restriction sites. Double underline 
designates indicated codons. Sequences not shown for rat aldolase B cDNA and 
pcDNA3™ vector are available in Tsutsumi, et al. (Tsutsumi et al., 1984) and from 
Invitrogen Technical Services, respectively, c) Amino acid sequence predicted for the 
carboxyl terminus of the RABM protein expressed from this vector (residues not shown 
would be identical to rat aldolase B), single letter amino acid code is underlined for 
residues added to create a 9E10 epitope. 

Site-Directed Mutagenesis of pcDNA3RABM 

I used three different strategies for PCR mutagenesis: (1) overlap extension (Ho 
et al., 1989), (2) Quick-Change™ Site-Directed Mutagenesis Kit (Stratagene #200518), 
or (3) restriction-limited insertion as described below (Fig. 2-4; Table 2-1). 

I used an overlap extension protocol adapted by Brian Cain from Ho and others 
(Ho et al., 1989). In brief, a mutagenic primer pair (see Fig. 2-3, positions 2 & 3, 4 & 5, 
or 6 & 7) was made complementary to each other and to base pairs on either side of a 
targeted change (non-complementary) in aldolase B coding sequence (Table 2-1). For 
example, Q58 mutagenesis started with pcDNA3RABM as template in two PCR 
reactions using primers at positions 1 & 2 (Fig. 2-3) to make product for the 5' end of an 
insert and at positions 3 & 8 (Fig. 2-3) to make product for the 3' end of an insert. PCR 
was done for 35 cycles (1 min. at 94°C, 1 min. at 55°C, and 2 min. at 72°C). 

At one end of one product, sequence was derived from primer at position 2 and 
therefore was reverse complementary to one end of the other product derived from 
primer at position 3. These products were combined in Taq polymerase buffer 
(Promega), melted at 94°C, and cooled very slowly to allow promiscuous annealing and 
yielding a small amount of 3 'end-to-end annealed single stranded DNA from 5' and 3' 
products for ends of a desired insert. Taq polymerase was added and the temperature 
raised to 72°C for run-off extension through ends complementary to positions 1 & 8 
(Fig. 2-3). Primers for positions 1 & 8 were added and cycled through temperatures as 
done previously which produced a smear of products due to promiscuous annealing. 


Rat Aldolase B 





FIGURE 2-3: Positions of Primers for Site-Directed Mutagenesis. Map shows the RABM coding region 
of pcDNA3RABM; positions drawn to scale. Numbered arrows identify 5'-> 3' primer sequences at 
complementary sites. Shaded areas are cDNA sequences for the indicated polypeptides. Letters B, E, 
and Xb indicate restriction sites for enzymes Bsal, EcoR V, and Xba I, respectively. E and Xb are sites 
of insertion into pcDNA3 (Invitrogen) multicloning site. 

TABLE 2-1: Details of Primers for Site-Directed Mutagenesis 



Position (bp) 


1 st Base of 

Start Codon 




Primer Sequence, 5 'to 3' 

(sequence non-complementary 

to pcDNA3RABM is underlined) 








+25 to 






+25 to 






+163 to 





+163 to 





+ 165 to 






+196 to 





+322 to 





+322 to 





+313 to 






+313 to 





+617 to 



*Fig. 2-3;**amino acid (single letter code) change by 
complementary to paired primer with position shifted 

residue # (start M = 1); ***primer non- 
for restriction-limited insertion (Fig. 2-4). 

Though extraneous products were common, the most abundant product was the desired 

mutated DNA fragment representing full length mutated insert. This insert containing 
altered DNA code was Bsal digested, gel purified, and ligated into the corresponding 
site of fresh pcDNA3RABM from which the wild type fragment was removed. 
Normally, restriction sites for primer positions 1 & 8 would be designed for two unique- 
site enzymes producing different overhangs. However, Bsal cuts outside its recognition 
sequence producing randomly unique overhangs that abrogate a need for separate 
enzymes. Bsal cuts a third site near the ampicillin-resi stance gene of pcDNA3RABM, 
producing two fragments of vector besides the insert fragment. Overhangs for all three 
sites were randomly different allowing three-fragment ligation with proper orientations. 

The Quick-Change™ Site-Directed Mutagenesis Kit (Stratagene) was also used 
and found to be more rapid. The manufacturer's protocols were followed, and primers 
designed for primer extension and amplification at Ql 1 1 (WID23 and WTD22 at 
positions 6 & 7) did not work with the Quick-Change™ kit. However, longer primers 
(WTD36 and WID35) were successful. The Quick-Change protocol involves in vitro 
synthesis of the entire vector (6.5 kb), possibly introducing errors anywhere in mutated 
pcDNARABM. To reduce sequencing, the Bsal fragment containing new mutations was 
cassetted into fresh vectors. All altered regions of vectors were sequenced at least twice 
to confirm changes in amino acid coding were specific for targeted residues. 

For Q58N mutation, a restriction-limited insertion was designed(Fig.2-4). This 
method uses a mutagenic primer (WID44) to span the Q58 codon and code for unique 

blunt-end restriction site, Seal (restriction-limited). Another primer (WID43) was 

designed with another unique blunt site, Pmll, such that Seal to Pml I blunt ends ligated 

to make proper aldolase B code. Steps were followed as indicated in Figure 2-4 to 

produce an expression vector for the mutated RABM, pcDNA3RABMQ58N. 

Expressing Epitope-ta g ged Aldolase B in Cell Lines 

Permanent Lines Expressing RABM 

E36 and ts20 cells were transiently transfected with pcDNA3RABM, using Lipofectin 

(GIBCO BRL) or DOTAP (Boehringer) by the manufacturers' protocol. When 

transfected cultures approached confluency labeled cells occurred in groups of 2 to 8 

presumably due to cell division. Transfected cultures were trypsinized at confluency and 

diluted > 1 : 1 5 by area into G418 Medium. Fresh G418 Medium was provided every 1-2 

days as needed to remove cell debris and maintain strong selection. By 2-3 weeks post 

plating, colonies of resistant cells were isolated, passaged, and screened for RABM 

expression (immunofluorescence microscopy and western blotting with anti-9E10 

monoclonal antibody as described below). 

Different lines permanently expressed RABM at varied levels (~10 fold range on 

western blots). Using experimental protocols described below, no effect was seen with 

doubling time, degradation of RABM, degradation of total protein, or viability. Of eight 

positive clones that continue to express after culture amplification, seven (4 from E36 

and 3 from ts20) maintained stable relative levels of RABM for twenty additional 

continuous doublings or longer (by immunofluorescence and on western blots, data not 


target codon 
for Q58 


C G 


Xho I 


A G A 

C A 






WID44 / GA G A 

Seal 5' 


1. PCR with WID5 and WID44 

2. Digest with Xho 1 and Sea I 



1. PCR with WID44 and WID9 

2. Digest with Pml I and Xba I 







3. Ligate to make pcDNA3RABMQ58N 


pcDNA3RABM digested with Xho I and Xba I 

Figure 2-4: Restriction-Limited Insertion for Constructing pcDNA3RABM058N . 
Above, primers (small arrows and sequences for WID43 and WID44) are shown relative 
to pcDNA3RABM template (bar). Sequences juxtaposed to template indicate 
complementary regions. Relative positions of Q58 codon and restriction sites are 
indicated. Below, DNA pieces for a three-fragment ligation. See text, Fig. 2-3, and 
Table 2-1. 

shown). To facilitate experimental quantification, the highest RABM-expressing E36 

and ts20 cell lines fully designated E36RABM14.1 and ts20RABM10.2 or abbreviated 

throughout this dissertation as E36AB and ts20AB, respectively, were used. In control 

experiments, variation in the level of RABM expression did not affect results (data not 


Similar procedures were repeated with Fao rat hepatoma cells transfected, except 
another plasmid pHAHAB also was used to express human aldolase B tagged with the 
12CA5 HA epitope at its amino terminus (HAHAB). Screening for HAHAB expression 
was done with monoclonal antibody against 12CA5. When G4 1 8-resistant clones were 
isolated at very most 20% of cells in a given clone had visible expression that was mostly 
dim with a few bright cells, and this fraction was rapidly lost with culture splitting for 
amplification. Subcloning and screening of a few hundred colonies produced one truly 
stable line expressing HAHAB at levels comparable to ts20AB expression of RABM. 
This line was designated FaoAB. FaoAB cells split at low density (1 :20 or less) grew 
much slower than parent Fao cells. In passages using about 1 : 10 or 1 : 15 splits, growth 
rates were similar to parent Fao cells. An attempt to isolate HuH7 cell lines expressing 
RABM was made, but failed to produce any clones having permanent expression. 
Transient Transfection System 

Fao, HuH7, NRK, and BHK (hepatic and renal cell lines derived from tissues that 
express aldolase B) were transfected for transient RABM expression with a series of 
lipids according to manufacturer's standard protocols (Invitrogen and Boehringer- 

Mannheim). Transient transfection gave very broad cell-to-cell variation in expression by 

9E10 immunofluorescence, but ratios of bright to dim cells were relatively reproducible 

between transfections. Transfection efficiency was defined as fraction of labeled cells. 


Cells were grown on glass coverslips to desired confluency, rinsed briefly with 
PBS and fixed with 4% paraformaldehyde in PBS for 20-30 minutes. Fixed cells were 
washed three times for 10 minutes in 50 mM ammonium chloride/ 0.1% Tx-100/PBS. 
Coverslips were placed on drops containing antibody diluted 1 : 100 in 5%NGS/ 0. 1% 
Tx-100/PBS for 1-2 hours at room temperature. Coverslips were washed four times for 
5 minutes in 0. 1% Tx-100/PBS. Coverslips were placed on drops containing an 
appropriate secondary antibody (rhodamine or fluorescein conjugated) diluted 1 : 100 in 
5%NGS/ 0.1% Tx-100/PBS for 1 hour at room temperature. Coverslips were washed 
six times for 5 minutes in 0.1% Tx-100/PBS, then mounted on Fluoromount G (GIBCO 

Preparation of Ubiquitin-Free Aldolase B Antigen 

Different E. coli strains were transformed with pXPB, a plasmid vector kindly 
provided by Dean R. Tolan for bacterial expression of enzymatically active human 
aldolase B (Beernink and Tolan, 1992). The aldolase B expressed in E. coli retains all 
the enzymatic properties of the protein isolated from human liver (Sakakibara et al., 

1989). Increased expression of a 40 kD Coomassie signal in SDS-PAGE of whole cell 

preparations was apparent in transformed cells. 

E. coli JM83 cells were transformed with pXPB, which produced much more 40 

kD protein than untransformed cells, about 0.3 mg per ml of 1.9 OD 6 oo nm culture. A 250 

ml culture of LB broth + ampicillin (100u,g/ml) was grown to 1.9 OD 6 oonm culture and 

pelleted in a Beckman GSA rotor at 3500 rpm for 10 min. Samples were maintained at 

0-4°C for the rest of the procedure. The cell pellet was resuspended in 20 ml 15% 

sucrose/50 mM EDTA/50 mM Tris-HCl pH 8.5. To this, 5 ml 5 mg/ml lysozyme was 

added, gently mixed by inverting, and incubated 15 min. Then 15 ml 0.1% Triton X- 

100/50 mM Tris-HCl pH 8.5 was added, gently mixed, and incubated with periodic 

inverting for 20 min. After centrifugation at 9000 rpm for 30 min. (GSA rotor), the 

supernatant was decanted into Polyclear tubes and ultracentrifuged in aSW27 rotor at 

23,000 rpm for 60 min. The resulting supernatant was the crude extract which was 

further processed as previously described for isolation of aldolase B from liver extracts 

(Penhoet and Rutter, 1975). Saturated (NHO2SO4 was slowly added (0.5 ml/min.) to 

45% final concentration with constant stirring. After centrifugation at 9000 rpm for 60 

min., the supernatant was collected and the pellet discarded. (NtL^SCu was slowly 

added to 60% concentration, and 6 N NH4OH added to pH 7.5. The mixture was 

immediately transferred to centrifuge bottles and let stand for >2 hours. After 

centrifugation at 9000 rpm for 60 min., the 60% (NH^SCu pellet was dissolved in 1 

mM EDTA/10 mM Tris-HCl pH 7.5 and dialyzed against the same buffer. The sample 

was loaded onto a 25 ml phosphocellulose (fine mesh, 1.26 meq/g) column prepared 

exactly according to Penhoet and Rutter. The column was washed with 5 mM EDTA/50 

mM Tris-HCl pH 7.5 (50-60 ml) until OD 2 8o nm approached zero. Then 2.5 mM fructose 

1,6- diphosphate in 1 mM EDTA/10 mM Tris-HCl pH 7.5 was used to specifically eluted 

a sharp aldolase peak. Peak fractions with specific activities of 0.83 to 0.98 aldolase 

U/mg (1 U/mg expected for aldolase B) were pooled, precipitated by 55% (NEL^SC^, 

and stored as a suspension at 4°C. Before immunization, the suspension was dialyzed 

into 10 mM Tris-HCl pH 7.5. According to Coomassie labeled SDS-PAGE and 

enzymatic properties (Rutter et al., 1966), the recombinant human aldolase B constituted 

at least 95% of the final protein and was more than 99.99% pure of bacterial aldolase 

activities, containing <0.005% bacterial isoform (EDTA-sensitive) activity. 

Production of Antibodies Against Aldolase B 

Rabbits were fed and housed by University of Florida Laboratory Animal 

Services. Antibodies to native and denatured aldolase B were raised as previously 

described (Reznick et al., 1985) except ubiquitin-free human aldolase B antigen was used 

(as prepared above). For making antibody to native aldolase B, 50 ug antigen in 0.5 ml 

10 mM Tris-HCl pH 7.5 was emulsified 1:2 with complete and incomplete Freund's 

adjuvants for immunizations and boosts, respectively. For antibody to denature antigen, 

the antigen solution was supplemented with 2%SDS/2%(3ME and boiled 10 min. prior to 

mixing with adjuvant and immediately before administration to animals. Intradermal 

injections were in the thoracic region on the backs of specific pathogen-free New 

Zealand White rabbits. The first boost was two weeks after initial immunization. 

Preparative Western blots of rat liver cytosol were routinely used to follow 
specific immunoreactivities. One week after the first boost, the rabbit receiving native 
antigen produced a highly reactive serum specific for 40 kD aldolase B which was 
maintained without further boosts. Unless otherwise specified, these antibodies were 
used to detect native aldolase B-specific polyclonal epitopes throughout this study. 

One week after the first boost, the rabbit receiving denatured antigen produced 
antibody that specifically labeled a 68 kD protein moderately, 60 kD and 78 kD proteins 
lightly, and a high molecular weight smear. This pattern was remarkably similar to that 
for anti-Ub68. Furthermore, relative recognition of isoforms A and B were comparable 
to that for Ub68. Thus, denaturation of aldolase B disrupted isoform-specific epitopes 
and produced anti-denatured aldolase B that similarly recognized epitopes in both 
aldolases A and B. Interestingly, the early bleeds had little or no reactivity to 40 kD 

Every 4-7 weeks (after injection sites completely healed), the rabbit immunized 
with denatured aldolase B was boosted. After the second boost, antibody 
immunoreactivities were greatly increased. Though these sera recognized some non- 
protein epitopes, they retained specificity for proteins labeled with sera from earlier 
bleeds. After the third boost, sera recognized 40 kD aldolase B but never as well as 
antibody to native antigen. 

Other Antibodies 

Anti-Ub68 was provided by William A. Dunn, Jr. and is a polyclonal rabbit 
antibody raised against a major ubiquitin-protein conjugate purified from lysosomes 
(Lenk et al., Submitted 1998). Monoclonal mouse antibodies against thel2CA5 HA and 
9E10 myc epitopes were obtained from the University of Florida Hybridoma Core 
Laboratory. Alternatively, hybridoma cells expressing anti-9E10 (provided by the 
Hybridoma Core) were injected into the peritoneum of BALB/c mice (provided by the 
University of Florida Laboratory Animal Services Division), and periodically, ascites was 
harvested until mice showed signs of discomfort or disease at which time they were 

Viability Assays 

All treatments were measured for viability except for incubations in HBSS 
(Hank's Balanced Salts Solution) which were used to induce cell death. Heat stressed 
cultures using HBSS had many cells rounding and sloughing off culture surfaces in 4 to 6 
hours of treatment, and were only used in experiments to determine the effect of dying 
cells on protein degradation measurements. When HBSS was replaced with MEM 
(Sigma #M-0268 + 2.2 g/1 sodium bicarbonate), such rounding and sloughing was 
delayed beyond 25 hours of treatment. To measure metabolic viability, cultures exposed 
to each experimental treatment were recovered under normal maintenance conditions for 
12 hours followed by addition of the same labeling medium (0.1 mCi 35 S-methionine/ml) 
used for protein degradation assays. Then protein synthesis was measured as TCA 

precipitable counts incorporated in 20 min. Metabolic viability was defined as the % 

protein synthesis relative to duplicate cultures treated with fresh medium under normal 

maintenance conditions. Since experimental treatments (see Experimental Conditions) 

lack serum and nutrients (MEM instead of DMEM) relative to maintenance conditions, 

our assay probably underestimates metabolic viability. In MEM, more than 80% of 

metabolic viability was retained for 28 hours and 15 hours in heat stressed E36 and ts20 

cells, respectively. Unless otherwise stated, all data are reported for incubations and 

drug treatments that retained 90% or greater metabolic viability. For low density 

cultures (<10% confluent), the Cell Titer 96 AQ System® was used according to 

manufacturer's protocols (Promega) which indirectly measures electron transport 

pathway activity. 

Subcellular Fractionation 

Subcellular fractionation of E36 and ts20 cells is summarized in Figure 2-6. To 
produce a homogenate containing intact organelles, cells were grown to recent 
confluency in 100 mm dishes, using 10 ml DMEM (Sigma #D-5648) + 10% FBS 
medium. Culture rinsed with 10 ml DPBS (Sigma D-5652) was treated with 1.0 ml IX 
Trypsin-EDTA (Sigma #T-9395 diluted in DPBS),and incubated at room temperature 
until cells easily and completely knocked loose from the plate (10 minute maximum). 
After adding, 5 to 10 ml medium containing 10% FBS, cells were transferred to 15 ml 
conical centrifuge tubes (polypropylene), centrifuged at 1,500 X g for 5 minutes, and 
supernatant discarded. Pellet was suspended in 1.0 ml ice-cold cavitation buffer (SHE): 

Figure 2-6: Subcellular Fractionation Scheme for E36 and ts20 Cells . 50 
Cell Culture 


Scrape cell sheet 


Centrifuge 1000 xg v c 

~~ r Supernatant = bC 




Homogenize __v T T 

~ r Homogenate = rlO 

Centrifuge 1000 xg _^ peUet 

Supernatant T „ 

F Pool = LP 

4 t 

Centrifuge 1000 xg _^ peUet 


Supernatant = LS 


Centrifuge 100,000 xg _^ Supernatant = HS 






Centrifuge 100,000 xg _> ^ _ Hp 


Supernatant (wash , dilute HS) — > <1% of culture content 

250 raM sucrose, 10 mM HEPES (Research Organics #6003H-3), 1 mM EDTA, pH 7.4, 

loaded into a N 2 -cavitation bomb chamber using 2-3 ml total SHE volume, pressurized 

to 65 psi for 10 minutes, and collected sample from bomb directly into a pre-chilled 

Dounce homogenizer. Procedural details were as described by the cavitation bomb 

manufacturer's specifications (Kontes). The sample was homogenized with 5 strokes of 

a pestle, minimizing froth by limiting passage of bubbles to the sample side of pestle. 

This was saved as the homogenate (Ho). 

Alternatively, the trypsinization step was replaced by scraping the cell sheet 
directly into SHE which resulted in a large fraction of cytosol but not organelles to leak 
out. This facilitated the separation of aldolase associated with organelles from soluble 
aldolase in the cytosol. Scraped cells were pelleted as done above to remove trypsin 
solution, but in this case the supernatant was saved (to assess cytosolic leakage) as the 
scrape fraction (Sc). The rest of cavitation and homogenization was performed as 

Homogenate (Ho) was fractionated essentially as described previously 
(Rickwood, 1992; Coligan et al., 1995), using Sigma reagents for assays and Sorvall or 
Beckmann centrifuges and accessories. Homogenate was centrifuged at 1,500 X g for 
20 minutes yielding a low speed pellet (LP) and supernatant. The centrifugation was 
repeated with the supernatant to make sure nuclei, large debris, and unbroken cells were 
efficiently removed, the resulting pellet was pooled in LP, and the resulting low speed 
supernatant (LS) was centrifuged at 100,000 X g for 90 minutes, yielding a high speed 

supernatant (HS) and pellet. The pellet was resuspended in well over 400 volumes of 

fresh SHE and centrifuged at 100,000 X g for 60 minutes. The supernatant had a 

content similar to HS but was much more dilute (data not shown), so it was not pooled 

with HS. The pellet was saved as the high speed pellet (HP). 

Fractionation conditions were developed to separate abundant cytosolic aldolase 
from that associated with organelles. An initial scrape into fractionation buffer caused a 
three fold greater leakage of aldolase than acid phosphatase, so this step was retained in 
the procedure. Conditions were chosen to maximize recovery of lysosomal organelles 
(acid phosphatase and (3-hexosaminidase) from LP to HP and minimize release of 
lysosomal markers into Sc and HS. Recovery was reasonable (85-99% accounted). 

Enzyme Assays 

Aldolase assays were performed as previously described (Penhoet and Rutter, 
1975). "Aldolase reaction mix" includes 50 ul 6.3 mg/ml a-glycerophosphate 
dehydrogenase-triose phosphate isomerase mixture + 4 mg NADH (99% pure, Sigma) + 
20 ml 0.1 M glycylglycine pH 7.5. Add 5-50 ul of sample to 1 ml of aldolase reaction 
mix, and measure background AOD (340 nm)/min. (BG); add 50 ul 50 mM fructose- 1,6- 
diphosphate (FDP) or 100u,l 100 mM fructose- 1 -phosphate (F-l-P), and measure assay 
AOD (340 nm)/min. (ASSAY). Aldolase activity in units, U - (ASSAY-BG)/12.44 for 
FDP or - (ASSAY-BG)/6.22 for F-l-P. 

Acid Phosphatase 

Acid Phosphatase activity was the OD (405 nm) in lOOOx g supernatant after 60 
minute incubation at 37°C for 50 ul sample in 200 ul 8 mM p-nitrophenolphosphate/ 
2mM MgCl 2 / 90 mM Na-acetate pH 5.0 stopped by 600 ul 0.25 M NaOH (Rickwood, 

Protein Analysis 

Sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE), gel 
staining and drying, western blotting, and autoradiography followed standard protocols 
similar to those described and referenced elsewhere (Ausubel et al., 1994; Coligan et al., 
1995). Gels were made with 10% polyacrylamide (1:35 bis-acrylamide). 
Immunochemiluminescent detection was done with Amersham ECL Reagents and 
protocols, except blots were rinsed in Kodak 1XCDS buffer just before exposure to ECL 
chemicals. Protein concentrations were determined with Bio-Rad protein assay (IgG 
standard) reagent or spectrophotometric absorbence (205 nm or 280 nm). 

For direct immunoprecipitations, monoclonal antibody was purified and 
conjugated to sepharose 4b by a CNBr protocol similar to that described in Coligan et 
al., 1995. Cells were incubated on a shaker at 0-4°C with ice cold minimum lysis buffer 
(MLB: 1%NP40/ 1 mM EDTA/50mM Tris-HCl pH 7.4), standard lysis buffer (SLB: 
150 mM NaCl/ MLB), or modified radioimmunoprecipitation assay buffer (mRIPA: 
0. 1% SDS/ 0.25% desoxycholate/ SLB) supplemented with a cocktail of protease 
inhibitors (leupeptin, TLCK, pepstatin, aprotinin, and PMSF obtained from Sigma or 

Boehringer/Mannheim) at concentrations according to Harlowe. The mRJPA gave the 

best results with mAb's conjugated to sepharose, and SLB gave the best results with 

polyclonal sera precipitated with commercial protein A-agarose (Boehringer-Mannheim). 

Lysates were precleared for 15-20 minutes on maximum in a microfuge then transferred 

to tubes containing 4-10 ul packed bed volume of mAb-sepharose or 1-5 p.1 antiserum 

then rotated in the cold. After one hour, tubes containing antiserum only received 10 ul 

packed bed of protein A-agarose. Rotating incubation was continued for varying times, 

usually overnight, and such times were held constant within a given experiment. Rapid 

washes by quick microcentrifugation and aspiration of supernatants were done three 

times with the same lysis buffer and one time with TBS, then immediately processed for 

other procedures. For SDS-PAGE, a 2-fold concentration of sample buffer was directly 

applied to the pellet and heated at 95-100°C for 5-10 minutes. For precipitable aldolase 

activity, pellets were further washed with TE pH 7.5 then resuspended in aldolase 

reaction buffer. 

Densitometry (autoradiographs, chemiluminographs, & Coomassie gels) were 

quantified using a desktop scanner and Sigmagel software. Automatic brightness and 

contrast settings determined initial settings that were then kept constant. Spot settings 

were chosen to encompass regions of interest, reduce background, and optimize signal. 

Standard curves were routinely performed to characterize linear response ranges for 

relative protein levels which also defined backgrounds and allowed quantification of 

relative signals. Where protein bands were specific for transfected cell lines, 

untransfected cells were run in parallel through procedures and quantified to establish 


Stress-Induction of Protein Degradation 

Culture Preparation 

Cells were plated, grown, and maintained at confluency for 2-3 days. About 30- 

40 hours before an experiment, cultures were fed with standard maintenance medium or 

for protein degradation assays, 14 C-Valine or 35 S-Methionine was made up in comparable 

medium with the corresponding unlabeled amino acid omitted. At the beginning of an 

experiment, cells were switched to the media and temperatures indicated in Table 2-2. 

Table 2-2: Comparison of Systems for Stress-Induced Degradation of Proteins 

Heat Stress Induction 

Starvation Induction 















*KH, Krebs-Heinseleit medium (Lefer et al., 1982) 
Heat Stress 

Cultures were prepared as described above and replicates were incubated under 
control temperature (CT) and heat stressed (HS) conditions (Table 2-3) which are 
permissive and non-permissive, respectively, for the ubiquitin-activating enzyme El of 
ts20 cells (Handley-Gearhart et al., 1994; Kulka et al., 1988; Lenk et al., 1992). 
Accordingly, CT conditions included 4 mM bicarbonate-buffered MEM under 5% C0 2 at 
30.5°C, and HS conditions included 20 mM HEPES-buffered MEM under air at 41.5°C 
for 1 hour followed by 39.5°C. These incubations followed established protocols for the 


E36/ts20 cell system, except MEM replaced HBSS to improve viability. In some 
experiments, media were supplemented with 5-10 mM 3-methyladenine (3MA) or 40- 
160 uM chloroquine (CHQ) as indicated in Results. Logistics and consistency with 
published protocols required differences in buffering and atmosphere between CT and 
HS conditions. As a control for such 

Table 2-3: Heat Stress (HS) and Control Temperature (CT) Treatments 
to Determine Ubiquitin-Activating El Mediated Processes 


MEM buffer* 





2.2 g l\ NaHC0 3 


5% C0 2 



20 mM HEPES 

1 hour@41.5°C 
then 39.5°C 

in water bath 

Repeat treatments with two cell lines: E36 (parent) and ts20 (mutant) 

*For protein degradation experiments, medium was supplemented with unlabeled amino 
acid corresponding to that used for labeling (5 mM methionine or 10 mM valine), 
differences, CT as summarized in Table 2-3 was compared with CT in HEPES-buffered 

MEM under air, yielding no differences in cell morphology, viability, or protein 

degradation measurements (data not shown). The other control (i.e. comparing HS in 

conditions above with HS in bicarbonate-buffered MEM under 5% C0 2 ) was not tested, 

because bicarbonate buffering varies with temperature. 

Starvation (Nutrient Stress) 

Cultures were prepared as described above and replicates were refed with fresh 

standard maintenance medium (DMEM + 10 % FBS) or Krebs-Heinseleit (KH) medium 

and referred to as "Fed" or "Starved," respectively. KH components are given in Lefer, 

et al., 1982. For protein degradation experiments, media were supplemented with 

unlabeled amino acid corresponding to that used for labeling (5 mM methionine or 10 

mM valine). Inhibitor treatments utilized the same levels as for heat stress above. 

Protein Degradation 

Permanent cell lines or transiently transfected cultures were treated according to 
instructions in the section Stress-Induction of Protein Degradation to produce cells with 
metabolically labeled proteins containing 35 S-methionine or 14 C-valine. Cell sheets were 
routinely rinsed with DPBS (Sigma) immediately followed by application of media 
containing unlabeled excess amino acid (5 mM methionine or 10 mM valine) to cultures 
described above. This initiated chase of radiolabel incorporated into proteins. 

At various times, aliquots of media were collected and TCA precipitated to 
measure release of soluble counts measured with a scintillation counter. At the end of 
the chase, whole cultures were TCA precipitated to determine total counts. Fraction of 
TCA soluble counts released at various times were subtracted from 1 to calculate TCA 
precipitable counts, representing the remaining total radiolabeled protein at those times. 

Alternatively, cultures were harvested at each time point, processed for 
immunoprecipitation, SDS-PAGE, and autoradiography, and the radioactive signals in 
specific protein bands (absent in untransfected cells) were quantified by densitometry, 
representing the remaining radiolabeled protein (aldolase) at those times. 

To compare degradative rates for proteins expressed at different concentrations, 

relative rates are normalized to the initial amount of the protein, giving a fractional 

change per time with units %/hour. This fractional rate is constant for unchanging 

degradative mechanisms regardless of a substrate protein's concentration. For a given 

protein, this fractional rate defines the degradative rate constant, kd. For total proteins, 

the fractional rate represents weighted average kd contributed by the varied amounts of 

different proteins. Throughout this dissertation, all degradative rates and other rates of 

protein decrease are estimated with the following calculation. Fractions of radiolabeled 

protein remaining were transformed by the natural log (In) and regression analysis 

performed using the following function: 

ln(100*S t /S ) = -k d • t, 

where S t = signal at time t, S = initial signal, kd = first order degradative rate constant, 
and t = time. Degradation rates were taken as the negative slope of the regression (kd) 
and the standard error of the slope was calculated as the standard error of y at x divided 
by the square root of the deviations squared of x. Degradative turnover is also described 
by half-life, t Vl , the time needed to replace 50% of existing molecules with new ones. In 
general, the degradative rate constant and half-life are simply converted by kd = ln(2)/ t Vl 
= 0.693/ 1. /2 . Note that some investigators do not follow the empirically confirmed first 
order relationship of kd and t*. This results in kd and t> /2 reported in the literature that can 
vary by as much as three fold from similar data presented using conventional 



Aldolase B undergoes degradation during stress via autophagy. In Chapter 1, 
evidence was described for ubiquitin protein conjugates which underwent starvation- 
induced enrichment in autophagic vacuoles and lysosomes, and preliminary work 
suggested that putative ubiquitinated aldolase B was amongst these conjugates (Figs. 1-3 
and 1-4). This resulted in the hypothesis that stress-induced autophagic degradation of 
aldolase B requires ubiquitination. In the first part of this chapter, the presence of 
aldolase B in ubiquitin conjugates is confirmed, and a role for ubiquitination in the heat 
stress-induced autophagic degradation of aldolase B is demonstrated. 

Two mechanisms were found to simultaneously mediate enhanced degradation of 
aldolase B during heat stress: autophagy and cytosolic proteolysis. To separately 
examine effects on autophagy, heat stress-induced changes in endogenous aldolase A and 
exogenous epitope-tagged aldolase B associated with pelletable organelles were assayed. 
In addition, autophagic degradation of long-lived proteins was demonstrated to require 
ubiquitination. The results described below support a role for ubiquitination in the 
function of a subset of lysosomal proteases. 


In Vivo Multiubiquitination of Aldolase B 

Previously, antibodies were raised to a major ubiquitin-conjugated protein, Ub68, 
that associated with autophagic vacuoles and lysosomes during stress-induced 
autophagy. On western blots of subcellular fractions isolated from rat liver, anti-Ub68 
produced a pattern suggestive of a 40 kD protein successively conjugated with serially 
increasing numbers of ubiquitin. Such a pattern is referred to as a ubiquitin ladder for 
the modified core protein. Peptide sequence analysis identified the 40 kD protein as 
aldolase B, suggesting that anti-Ub68 reactive proteins might represent a ubiquitin ladder 
for aldolase B. 

To confirm the existence of ubiquitinated forms of aldolase B, antibodies (Fig. 3- 
2) were raised against aldolase B (Fig. 3-1) and used to assay western blots of rat liver 
fractions from the previous studies (Fig. 3-3). In short, anti-aldolase B recognized the 
same proteins as anti-Ub68, confirming a ubiquitin ladder for aldolase B and 
demonstrating that aldolase B is multiubiquitinated in vivo. The data are consistent with 
ubiquitin-mediated autophagic degradation of aldolase B. 

First, antibodies against aldolase B had to be produced. However, aldolase B 
isolated from animal cells would need to be purified away from contaminating ubiquitin 
and ubiquitin conjugates. Since bacteria lack ubiquitin, aldolase B was expressed in E. 
coli and purified to produce ubiquitin-free antigen for immunization (Fig. 3-1). E. coli 
strains transformed with an expression vector for human aldolase B (Fig. 3- la, lanes x) 
produced more 40 kD protein (expected size of aldolase B) relative to untransformed 
cells (Fig. 3- la, 

Figure 3-1 : Isolation of Ubiquitin-Free Aldolase B Expressed in E. coli . a) Three E. coli 
strains DH5a, JM109, and JM83 were transformed for aldolase B expression as 
described in the text, pelleted, suspended in sample buffer, boiled, and run on SDS- 
PAGE with 60 ui culture (OD 6 oo nm = 1.5) equivalents per lane, u and x designate 
untransformed and transformed cells, respectively. The 40 kD band specific for 
transformed cells is indicated; b) E. coli JM83 cells expressing aldolase B were 
fractionated as described in the text: 1 and 2, whole cell preparations (as in part a) 
untransfected and transfected, respectively; 3, crude extract; 4, 45% (NH^SCU 
supernatant; 5, 45-60% (NH^SC^ cut; 6, aldolase activity peak from phosphocellulose 
column; 7, dialyzed antigen ready for immunization; 8, rat liver cytosol used to screen 
antibodies; 9, detection of lane 8 with antibody raised against protein in lane 7. Lanes 3 
to 6 were loaded with 12 ng, lane 7 with 4 ug, lane 8 with 25 ug, and lane 9 with 10 \xg 
of protein. Dark bands on light background indicate Coomassie R-250 label in gels, and 
light bands on dark indicate western blotted proteins detected with anti-native aldolase B 
by ECL (Amersham); c) Elution profile for phosophocellulose chromatography, Relative 
Amounts: Protein, OD(280 nm); Aldolase Activity, mU/10 (loaded then started washes 
when collecting fraction 9); FDP i, elution started with fructose 1,6-diphosphate. 


DH50C JM109 JM83 1 

CI. U X U X U X D. 1 

2345 6789 

40 ► 


< ► 










10 15 20 25 30 
Fraction Number 

35 40 45 

Figure 3-2: Antibodies Against Aldolase B . a) A preparative western blot of rat liver 
cytosol run on SDS-PAGE (see Coomassie in Fig. 3 -6b, lane 8) was prepared and strips 
containing approximately 3 \xg total protein were probed with sera from rabbits 
immunized with native and denatured aldolase B as indicated, each lane contains a strip, 
numbers above each lane correspond to bleed numbers, and numbers with arrowheads 
indicate molecular weight, b) A Bio-Rad slot-blotting apparatus was used to load 10 
ng/slot of antigen indicated by vertical labels (Aldolase A from Sigma), each row was 
probed with serum raised against antigen indicated by horizontal labels with increasing 
serum dilutions shown at the bottom of the figure, corresponding preimmune sera were 
used in rows immediately above and below anti-aldolase B and anti-Ub68, respectively. 





< a 


o a 



Aldolase B 




Aldolase B 



3200 12800 

Figure 3-3: Aldolase B Ubiquitinated In Vivo is Enriched in Lysosomes . a) Preparative 
SDS-PAGE of cytosol from starved rat liver was western blotted onto nitrocellulose 
then cut into strips with ~3 ug protein/strip, and stained with early antisera against 
aldolase B (Fig.3-2a, bleed 4) or Ub68 (bleed 5); b) Aldolase B antisera from bleeds 
after booster injections (Fig.3-6a, bleed 10) were reacted with Cy (cytosol strips as 
prepared above) or ML (similar strips using a lysosome-enriched fraction instead). N, 
antiserum to active native enzyme; N , same as N with 10-fold longer ECL exposure; D, 
antiserum to chemically denatured enzyme; I, antiserum to antigen extracted from 
polyacrylamide gel slices; and P, preimmune serum. Arrowheads, molecular weights in 
kD. Dots on rightmost edge indicate bands at molecular weights higher than expected 
for 40 kD aldolase B subunit. Susan E. Lenk provided subcellular fractions defined as 
follows: a 1,000 x g supernatant of rat liver homogenate was centrifuged at 6,000 x g; 
the resulting pellet was enriched in lysosomes and contained mitochondria (ML); the 
6,000 x g supernatant was centrifuged at 100,000 x g yielding a supernatant fraction 
referred to as cytosol (Cy). 



Aldolase B 

Aldolase B Ub68 

a. N D P I P 

N N N N D D 

1 . Cy ML Cy ML Cy ML 

68 ► 

40 ► 

68 ► 

40 ► 


lanes u). Aldolase B was purified from the most productive strain, JM83, by cellulose 

phosphate chromatography (Fig. 3-lc) of a 45-60% ammonium sulfate cut (Fig. 3-lb, 
lane 5) from crude lysate (Fig. 3-lb, lane 3). Cellulose phosphate chromatography 
separates aldolase B by substrate affinity at the enzyme's active site, allowing 
enzymatically active aldolase B to be specifically eluted with fructose 1,6-diphosphate, 
FDP (Fig. 3-lc, peak at fraction #30). Aldolase B in peak fractions was at least 95% 
pure based on Coomassie stained SDS-PAGE gels (Figure 3-lb, lane 6) and specific 
activities ranging 0.95-0.98 U/mg (pure aldolase B = 1 .0 U/mg). EDTA resistance of 
purified aldolase B activity indicated that contamination by class II bacterial aldolase was 
less than 0.005% (data not shown). 

Aldolase B antigen described above was used to raise antibodies against native 
and chemically denatured aldolase B as detailed in Materials and Methods (Chapter 2). 
Preparative Western blots of rat liver cytosol were routinely used to follow specific 
immunoreactivities (Fig. 3-2a). Native antigen produced a highly reactive serum specific 
for 40 kD aldolase B. Anti-native aldolase B demonstrated minimal cross-reactivity 
with aldolase A (Fig. 3-2b). However, this antibody aldolase B effectively recognized 
aldolase B from different animal species(Fig. 3-2, a. 40 kD rat aldolase B, b. purified 
human aldolase B). 

A rabbit immunized with denatured aldolase B produced antibodies (Fig. 3 -2a 
"Denatured" bleeds 2 through 5) that specifically labeled a pattern indistinguishable from 
that for anti-Ub68 on western blots of subcellular fractions from rat liver (compare Fig. 

3-3, lanes D with Figs 1-3 and 1-4, anti-Ub68). This demonstrated aldolase B epitopes 

in previously identified ubiquitin-protein conjugates and confirmed that aldolase B is 
ubiquitinated in vivo. The antibody to native aldolase B was specific for the 40 lcD 
unmodified monomer. However, with ten-fold greater exposure times, even the anti- 
native aldolase B detected a ubiquitin ladder (Fig. 3-3b, N ). Preimmune sera failed to 
label any proteins. Antibodies raised against native aldolase B are known to be highly 
specific (Haimoto et al., 1989). Given this, labeling with anti-native aldolase B provides 
even stronger evidence that aldolase B is multiubiquitinated in vivo, and confirms that 
Ub68 is probably a stable conjugate of the form: (aldolase B)i(Ub) 4 . 

Multiubiquitinated Aldolase B is Denatured and Enriched in Lysosomes 
Aldolase A and B can spontaneously refold to active enzyme after reversible 
denaturation treatments (Horecker, et al., 1972; Beernink and Tolan, 1996). Anti- 
denatured aldolase B preferentially recognized ubiquitinated (> 40 kD) forms, whereas 
anti-native aldolase B preferentially recognized the unmodified (= 40 kD) form (Figs. 3- 
2a and 3-3). These results indicated that ubiquitination inhibits spontaneous refolding of 
aldolase B into native conformations. Reznick and Gershon also raised antibodies 
against native and denatured aldolase B (Reznick et al., 1985) . Their anti-denatured 
aldolase B failed to immunoprecipitate catalytically active enzyme, but it effectively 
precipitated smaller peptides resulting from proteolysis. They also found that native 
aldolase B antibody failed to precipitate proteolytic fragments, but efficiently pelleted 
aldolase B activity (Reznick et al., 1985). The results reported in this study confirm the 

hypothesis that native epitopes of aldolase B require three-dimensional conformations 

that mask denatured epitopes. 

The 40 kD unmodified aldolase B predominantly occurred in cytosol (Cy) 

fractions with <10% in lysosome-enriched (ML) fractions (Fig. 3-3b). Consistent with 

previous results (Fig. 1-3), ubiquitin-aldolase B conjugates (bands >40 kD) were 

enriched in ML fractions with less occurring in Cy fractions (Fig. 3-3b). Taken together, 

the data suggest that ubiquitination can provide a mechanism for maintaining aldolase B 

in a denatured conformation. This could contribute to enhanced degradation of aldolase 

B by making degradative signals more accessible or by making the protein more 

vulnerable to proteases. 

Heat Stress-Induced Delivery of Aldolase A to Lysosomes Requires Ubiquitination 

Above, ubiquitinated aldolase B was confirmed to contribute to ubiquitin 

conjugates that are enriched in autophagic vacuoles and lysosomes during nutrient stress 

(Figs. 1-3, 1-4, and 3-3). The results suggested a role for ubiquitination in autophagic 

degradation of aldolase B. During heat stress, ubiquitin-dependent autophagic 

degradation of long-lived proteins has been demonstrated in E36 Chinese hamster lung 

cells (Gropper, et al., 1991; Handley-Gearhart, et al., 1994). Our results suggested that 

aldolase B was a possible substrate for this mechanism. However, E36 cells express 

endogenous aldolase A but not aldolase B (described later). Both aldolase A and B are 

established substrates for autophagy (reviewed in Chapter 1), so aldolase A was 

examined as a substrate for ubiquitin-mediated delivery to lysosomes during heat stress. 

In addition to autophagy, temperature-dependent cytosolic proteolysis 

contributes to increased protein degradation during heat stress (next chapter; Hough and 

Rechsteiner, 1984). During autophagy, cytosolic proteins, like aldolase A, are 

sequestered into organelles (autophagic vacuoles and lysosomes) that can be pelleted by 

differential centrifugation. To measure effects specific for the autophagic pathway, 

aldolase A activity associated with pelletable organelles was assayed. 

To examine a role for ubiquitination in the stress-induced degradation of aldolase 
A in lysosomes, a system developed by Schwartz and Ciechanover was utilized for 
measuring ubiquitin-dependent degradation of long-lived proteins (Gropper et al., 1991). 
During heat stress, E36 cells undergo enhanced autophagic degradation of long-lived 
proteins. However, ts20 cells derived from E36 cells harbor a temperature-sensitive 
mutation in ubiquitination. Heat stress is non-permissive for the mutation , so 
ubiquitination and enhanced autophagic degradation is inhibited in ts20 cells. The 
degradative phenotypes were confirmed and are presented at the end of this chapter. 

FDP aldolase activity was used to follow endogenous aldolase A, and acid 
phosphatase activity was used to follow organelles in subcellular fractions collected by 
differential centrifugation as described in Materials and Methods (Fig. 3-4). 
Fractionation was optimized to maximize organelles released from cells, indicated by loss 
of acid phosphatase from low-speed centrifugation pellets (LP), and to maximize 
organellar integrity, indicated by fraction of acid phosphatase retained in high-speed 
centrifugation pellets (HP). Acid phosphatase occurs as both an integral membrane 


a. Aldolase Activity 


E36 E36 ts20 


Cell Line and Treatment 


b. Acid Phosphatase Activity 

g 10% 


E36 E36 ts20 


Cell Line and Treatment 


Figure 3-4 : Ubiquitin-Dependent Association of Endogenous Aldolase A with 
Organelles. Aldolase (a.) and acid phosphatase (b.) reported as % total culture activity 
(mean ± SD, n = 3 cultures) for subcellular fractions collected from E36 and ts20 cells 
treated for 8.5 h as indicated; CT, control temperature; HS, heat stress; +CHQ, 80 uM 
chloroquine; subcellular fractions were collected (Materials and Methods) and are 
labeled only on the first set of three bars (E36, CT): LP, low-speed pellet (lOOOx g); HP, 
high-speed pellet (100,000x g); Sum, total pelleted fractions (LP + HP); different from 
CT, Student's t-test: *, p<0.06; **, p<0.03; ***, p <0.009; ****, p <0.0008. 

protein and as a soluble matrix protein inside lysosomes. For organelles isolated here, 

acid phosphatase activity (about 60% of HP) was released by freeze-thaw (data not 

shown), indicating that much of it was soluble in E36 cell lysosomes and served as an 

adequate indicator of organellar integrity. 

E36 cells were incubated at control (CT) and heat stress (HS) temperatures and 

subcellular fractions pelleted during differential centrifugation were characterized for 

aldolase and acid phosphatase enzymatic activities (Figs. 3-4 and 3-5). Relative to CT, 

HS treatment significantly increased aldolase activity distributed in pelletable fractions 

isolated from E36 cells (Fig. 3 -4a). During 8.5 hour incubations that were used, 

partitioning of aldolase to pelletable compartments had to be faster than loss. This is 

consistent with accumulation of nascent autophagic vacuoles peaking by 6 hours after 

autophagic induction (Lawrence and Brown, 1992). Chloroquine (+CHQ) caused a 

more significant increase in pelletable aldolase activity. The effect of chloroquine 

suggests that lysosomal degradation contributes to aldolase A flux out of pelletable 

organelles consistent with an autophagic mechanism. In support of this, accumulation 

caused by heat stress and chloroquine (Fig. 3-5, HS+CHQ) corresponds to a 1.5±0.3%/h 

(mean ± SD, n = 3) increase in the sequestration rate for aldolase A which was similar in 

magnitude to induced autophagic degradation for total long-lived proteins (data 

presented in a later section). The results support endogenous aldolase A of E36 cells 

undergoing autophagic delivery to lysosomes during heat stress. 


Aldolase Accumulation 


E36 E36 ts20 ts20 


Cell Line and Treatment 

Figure 3-5: Ubiquitination Mediates Lysosomal Accumulation of Aldolase A During 
Heat Stress . Using enzyme activities collected for Figure 3-4, aldolase was divided by 
acid phosphatase to indicate relative aldolase associated with organelles (mean ± SEM, n 
= 3); values were normalized to the 100,000 x g pellet (HP) of control temperature (CT) 
to reflect aldolase accumulation relative to unstressed conditions. Labels are as in Figure 
3-4. Student's t-test: **, p <0.03; ***, p <0.009. 

Heat stress (HS) by itself failed to affect the distribution acid phosphatase activity 
in subcellular fractions (Fig. 3-4b). This indicated that increase of aldolase A in pellets 
was not due to redistribution of lysosomal organelles and supported the idea that the 
aldolase A was undergoing enhanced accumulation. Chloroquine treatment (+CHQ) had 
no effect on total pelletable acid phosphatase (Sum) but caused a redistribution of acid 
phosphatase from high-speed pellets (HP) to low-speed pellets (LP). It is known that 

chloroquine treatment causes lysosomes to swell (Glaumann, et al., 1986). As a weak 

base, chloroquine accumulates in organelles proportional to their acidity, and mature 

lysosomes are the most acidic organelles. The results here indicate that some lysosomes 

became large enough to pellet at lower centrifugation speeds. 

A basic result of subcellular fractionation is that different pelleted fractions have 
different contents of organelles (Rickwood, 1992). To demonstrate that aldolase A 
accumulates in a subpopulation of organelles (presumably lysosomes), aldolase activity 
was normalized to acid phosphatase activity and calculated the accumulation of aldolase 
A in HP and LP relative to lysosomal content (Fig. 3-5). Significant accumulation of 
aldolase activity only occurred in HP fractions during heat stress. In the presence of 
chloroquine (+CHQ), there was a greater than two-fold accumulation of aldolase activity 
in HP fractions but not LP fractions. The results suggest that aldolase A containing 
organelles were preferentially isolated in HP even during CHQ treatment, and are 
consistent with heat stress causing accumulation of aldolase A in a subpopulation of 

A previous study has shown that the fractional volume of autophagic vacuoles 
and lysosomes does not significantly increase in heat stressed E36 cells (Lenk, et al., 
1992). Together, the data indicate that heat stress increases the flux of aldolase A into 
autophagic vacuoles during heat stress. Unlike wildtype E36 cells, heat stress-induced 
accumulation of aldolase activity with pelletable organelles failed to occur for mutant 
ts20 cells (Figs. 3-4 and 3-5). Since heat stress inhibits ubiquitination in ts20 cells, this 

suggested that aldolase A accumulation in organelles requires ubiquitination. The data 

support a role for ubiquitination in heat-stress induced sequestration of aldolase A. 

Using electron microscopic morphometry, a previous study shows that in heat stressed 

ts20 cells conversion of autophagic lysosomes into residual bodies is specifically 

inhibited, resulting in a 6-fold accumulation of lysosomal volume (Lenk, et al., 1992). 

The subcellular fractionation results here indicate that earlier events in autophagic 

degradation (aldolase A sequestration) might also involve ubiquitination. In conclusion, 

the endogenous aldolase A of E36 cells appears to require ubiquitination for heat-stress 

induced delivery to lysosomes. 

Heat Stress-Induced Lysosomal Proteolysis of Aldolase B Requires Ubiquitination 

Earlier in this chapter, ubiquitinated aldolase B in liver was shown to contribute 
to ubiquitin conjugates that are enriched in autophagic organelles during starvation- 
induced autophagy. Ubiquitination was required for heat stress-induced delivery of 
endogenous aldolase A to lysosomes of E36 cells, suggesting that aldolase A was 
degraded via ubiquitin-mediated autophagy. To examine whether aldolase B undergoes 
ubiquitin-mediated autophagy like aldolase A, subcellular fractionation studies in the last 
section were repeated with E36 and ts20 cells expressing epitope-tagged aldolase B 

E36 and ts20 cells were transfected and selected for permanent expression of rat 
aldolase B with the 9E10 myc epitope on its carboxyl terminus (RABM). The 9E10 
epitope allowed efficient immunoprecipitation needed for degradation assays and 


Figure 3-6: Transient Expression of RABM . a) E36 cells and b) ts20 cells transiently 
expressing rat aldolase B with a carboxyl terminal myc tag (RABM) were processed for 
immunofluorescence microscopy (Materials and Methods) and labeled with antibody 
against aldolase B (ALDB) or the myc epitope (9E10). Phase contrast (Ph) images for 
corresponding fields are shown below each immunofluorescent micrograph. Scale bar = 
50 uM. 

unambiguous identification of the exogenous aldolase B, RABM. In Figure 3-6, 
transient expression of RABM was easily detected in a small fraction of cells with 
antibody to either the myc epitope (9E10) or to native aldolase B (ALDB). Many 
unlabeled cells indicated that E36 and ts20 cells do not express endogenous aldolase B. 
Cell lines were isolated and screened for permanent RABM expression, and the highest 
expressing lines for E36 and ts20 cells were designated E36AB and ts20AB, respectively 
(Figs. 3-7 and 3-8). 

After clonal selection all cells in a microscopic field were labeled for RABM in 
permanent cell lines. Though most cells were brightly labeled, some were only dimly 
labeled. Such labeling remained constant after multiple culture passages and for different 
cell lines, suggesting that the variability was a trivial artifact of the immunofluorescence 
protocol. According to immunofluorescence and western blot assays, different cell lines 
had characteristic RABM levels that were maintained after multiple passages (data not 
shown). Control experiments performed with cell lines expressing 5 to 10 fold 
differences in RABM levels gave similar results. To facilitate detection, the highest 
expressing lines (E36AB and ts20AB) were used extensively, and data are reported for 
these lines. Immunofluorescent morphology indicated that RABM predominated in the 
cytosol as shown by the presence of negatively labeled nuclei and vacuoles, providing 
evidence that the recombinant protein demonstrated normal localization (Figs. 3-6, 3-7, 
and 3-8). 




Figure 3-7: Permanent Expression of RABM in E36 Cells . Transiently transfected E36 
cells were selected and screened for permanent RABM expression; the highest 
expressing cell line (E36AB) and untransfected cells (E36) were processed for 
immunofluorescent detection as described in Materials and Methods with anti-9E10 myc 
epitope (9E10, upper panels); phase contrast for corresponding fields are shown (Phase, 
lower panels) . Scale bar = 50 uM. 




Figure 3-8: Permanent Expression of RABM in ts20 Cells . Transiently transfected ts20 
cells were selected and screened for permanent RABM expression; the highest 
expressing cell line (ts20AB) and untransfected cells (ts20) were processed for 
immunofluorescent detection as described in Materials and Methods with anti-9E10 myc 
epitope (9E10, upper panels); phase contrast for corresponding fields are shown (Phase, 
lower panels) . Scale bar = 50 uM. 






CQ t 






Figure 3-9: Biochemical Detection of RABM Expression . Three confluent cultures (1, 
2, 3) for each of the indicated cell lines was trypsinized, pelleted, suspended in 2 x 
sample buffer, boiled, divided into two aliquots, run on duplicate SDS-PAGE gels, 
western blotted, and duplicated blots were stained for aldolase B (upper panel) or the 
9E10 myc epitope (lower panel); E36, original Chinese hamster lung cell line; ts20, an 
E36-derived ts-mutant in ubiquitination; E36AB; an E36-derived cell line permanently 
expressing RABM; ts20AB, a ts20-derived cell line permanently expressing RABM. 
Arrow heads mark molecular weights (kD) for a doublet detected with anti-aldolase B. 

On western blots of whole cells labeled with anti-aldolase B (Fig. 3-9), a single 

faint 40 kD band was detected in untransfected cells (E36 and ts20), consistent with 

cross-reactivity to aldolase A (Fig. 3-2b). Lysates of E36 and ts20 cells had aldolase 

cleavage activity 35 to 40 fold higher for fructose 1,6-diphosphate than for fructose- 1- 

phosphate (data not shown). This difference for the two aldolase substrates is 

characteristic for aldolase A, identifying this isoform as the prevalent endogenous 

aldolase of E36 cells; note, aldolase B was not detected by immunofluorescence (Fig. 3- 


For RABM-expressing cells (E36AB and ts20AB), a closely spaced doublet of 
bands was labeled with aldolase B antisera on western blots of whole cells (Fig. 3-9). 
One band occurred at 40 kD coincident with endogenous aldolase A of untransfected 
cells (E36 and ts20). A second strongly labeled band occurred at SDS-PAGE mobility 
corresponding to 41 .3 ± 0. 1 kD (mean ± SEM, n - 6). On duplicate blots labeled for the 
C-terminal myc epitope of RABM (9E10), long luminographic exposure times only 
showed the 4 1 .3 kD band (Fig. 3 -9). 4 1 .3 kD matched the predicted molecular weight of 
RABM, confirming that E36AB and ts20AB cells express full-length RABM. 

If differences between immunoreactivities for hamster aldolase A (E36 cell 
endogenous) and rat aldolase B (RABM) are similar to differences between titered 
immunoreactivities (Fig. 3-2b) for purified rabbit aldolase A (Sigma) and purified human 
aldolase B (Fig. 3-1), then immunoreactivities can be used to estimate RABM expression 
relative to endogenous aldolase A. Endogenous aldolase A immunoreactivity was below 

the linear range of western blot ECL assays in sample sizes subsaturating for aldolase B 

detection. Immunodetection in the low range variably underestimated aldolase A by 3 to 

>5 fold (data not shown), but allowed an upper limit for relative RABM expression to be 

estimated. Estimates varied in the range of 0.2-2.4 RABM per endogenous aldolase A. 

Though not precise, these overestimates indicated that RABM levels in permanent lines 

were equal or less than the endogenous. 

To determine whether RABM undergoes heat stress-induced autophagy, cultures 
of E36AB and ts20AB were incubated at control (CT) and heat stress (HS) temperatures 
and fractionated as done for untransfected cells in the previous section titled, "Heat 
Stress-Induced Autophagy of Aldolase A Requires Ubiquitination." Subcellular 
distributions for activities of aldolase A and acid phosphatase were indistinguishable 
between parental and RABM-expressing cell lines. The enzymatic activity for aldolase A 
is -10 fold more than aldolase B (Penhoet and Rutter, 1975). Given this, aldolase B 
enzymatic activity was too low to detect above endogenous aldolase expression in E36 
cells. To overcome this problem, aldolase B immunoreactivity was followed on western 
blots of subcellular fractions (Fig. 3-10). 

Figure 3- 10a compares the distribution of immunoreactivities for RABM 
(Aldolase B) and an integral membrane protein of lysosomes (LAMP2b) in subcellular 
fractions isolated from E36AB cells incubated at control temperatures (CT). Details of 
the subcellular fractionation are described in Materials and Methods. When cell sheets 
were scraped from culture dishes and pelleted, some soluble proteins leaked out of the 
cells, as indicated by the presence of RABM (Aldolase B) in the supernatant (Sc, Fig. 3- 

Figure 3-10: RABM Associated With Lvsosomes Undergoes Ubiquitin-Mediated 
Proteolysis . Confluent E36AB and ts20AB cell cultures were incubated for 8.5 total 
hours at indicated conditions and then fractionated as described in Materials and 
Methods; samples of fractions were separated by SDS-PAGE, western blotted and 
labeled for distribution of lysosomal membranes on higher molecular weight half of blots 
(upper panels) and RABM on lower half of blots (lower panels) using antibodies to an 
integral membrane protein of lysosomes (LAMP2b) or to native aldolase B (Aldolase B); 
a) E36AB cells incubated in CT (see below) conditions were used to show distributions 
for lysosomal membranes and RABM which were comparable for all treatments within 
the variability of ECL detection (Amersham), 1% total culture equivalent effractions 
was loaded per lane except HP (lane 6) which used 10%: Sc, supernatant from cells 
scraped then pelleted at 1000 x g; Ho, homogenate of lysed cell pellet; LP & Lsu, pellet 
& supernatant from 1000 x g of Ho; HP & Hsu, pellet & supernatant from 100,000 x g 
of Lsu; b) HP fractions isolated from cells incubated under indicated conditions were 
loaded for equal acid phosphatase activity to show relative content of full-length RABM 
(Aldolase B, 41.3), proteolyzed RABM (Aldolase B, 40.0), and lysosomal membranes 
(LAMP2b): Co, control temperature, in DMEM + 10% FBS (normal culturing 
conditions); CT, control temperature in MEM medium (experimental control); HS, heat 
stress in MEM medium; +CHQ, medium supplemented with 80 uM chloroquine. 

Simplified Diagram of Fractions*: 

Cell Sheet -> 


& 1000 xg-> 


sup = Sc 

pel, homogenize = Ho 

1000 xg-* 

sup = 

= Lsu 






pel = LP 


pel = 


^sup, supernatants; pel, pellets 


a. E36ABCT 

Sc Ho LP Lsu Hsu HP 



■ — i 


b. HP Fractions 







2 3 4 5 6 7 

*«, •• #* 9 


10a, lane 1). Membrane-bound organelles were retained in cells as indicated by the lack 

of LAMP2b label in Sc. 

The cell pellet was lysed, homogenized (Ho, Fig. 3- 10a, lane 2), and used to 

produce 1000 x g (low speed) pellet (LP) and supernatant (Lsu). Low speed pellets 

generally contain nuclei, large cell fragments, and unbroken cells (Rickwood, 1992). LP 

contained no detectable RABM, indicating that most cells were broken enough to lose 

soluble proteins to Lsu (Fig. 3-10, lanes 3 and 4). Supporting this, LP contained about 

15% of the aldolase A activity contained in Lsu (data not shown). The fact that some 

aldolase A (Fig. 3 -4a) but no RABM was detected in LP is probably due to the greater 

affinity of aldolase A for pelletable cell components compared to aldolase B (Kusakabe, 

et al., 1997). LAMP2b was approximately equally distributed in LP and Lsu, indicating 

that about half the organelles (at least lysosomes) cofractionated with large cell 

fragments or nuclei. In agreement with this, acid phosphatase activity was similarly 

distributed between LP and Lsu (data not shown). 

The Lsu was then used to produce 100,000 x g (high speed) pellet (HP) and 

supernatant (Hsu). Such high speed centrifugations pellet all membrane-bound 

organelles and leave soluble cytosolic (and leaked organellar) components in the 

supernatant (Rickwood, 1992). Consistent with this, all the detectable lysosomal 

membranes (Fig. 3- 10a, LAMP2b) in Lsu (lane 4) were pelleted out of Hsu (lane 5). To 

make RABM (Aldolase B) labeling in HP (Figure 3- 10a, lane 6) comparable to that 

loaded in lanes containing cytosol, 10 fold more HP equivalent was loaded. 

On western blots of directly harvested whole cells (E36AB and ts20AB), aldolase 

B immunoreactivity (RABM) primarily occurred at 41.3 kD (Fig. 3-9). In subcellular 

fractions, a large proportion (>40%) of aldolase B immunoreactivity occurred at 40 kD 

(Fig. 3-10a). This indicated that 41 .3 kD RABM was proteolyzed to ~40 kD size during 

fractionation. Like RABM, the LAMP2b-reactive protein was also proteolyzed as 

indicated by the presence of a smear below its band on western blots (Fig. 3-10a, 

especially visible in 10X loaded HP, lane 6). EDTA and storage on ice was used to 

reduce protein degradation in subcellular fractionations but was insufficient to prevent 

this presumably artifactual proteolysis. However, this result suggested that processing of 

RABM from 41.3 kD to 40 kD could be used as an indicator of proteolysis. 

Consistent for all treatments and fractions containing RABM (except HP), the 

artifactual proteolysis was limited to processing 40-60% of the RABM (Fig. 3-10a, lanes 

1, 2, 4, and 5). Only in HP fractions, processing of 41 .3 kD RABM to 40 kD was 

complete or nearly complete, such that aldolase B immunoreactivity collapsed from a 

doublet to a single band at 40 kD (Fig. 3- 10a, lane 6; Fig. 3- 10b, lanes 1, 2, 3, 5, and 6). 

However, if lysosomal degradation was inhibited by chloroquine (E36AB, HS+CHQ) or 

autophagic degradation blocked by the non-permissive ubiquitination of heat stressed 

ts20 cells (ts20AB, HS) then the complete processing of RABM in HP fractions was 

blocked, as indicated by the persistence of 41.3 kD RABM in a doublet (Fig. 3- 10b, 

lanes 4 and 7). These data indicated that RABM proteolysis occurred in lysosomes 

(CHQ-sensitive) and required ubiquitination (blocked in HS ts20AB). Whether greater 

proteolysis of RABM seen in HP occurred in cells or during fractionation was not 

determined. With the data from the last section, the results suggest that RABM 

(aldolase B), like endogenous aldolase A, utilizes lysosomal degradation that requires 

ubiquitination in E36 cells. 

To demonstrate the presence of ubiquitinated aldolase B in E36 cells expressing 

RABM, E36AB and ts20AB cell cultures were incubated at the different conditions used 

for subcellular fractionation studies above. Then RABM protein was isolated using 

9E10-specific immunoprecipitation and separation on SDS-PAGE (Materials and 

Methods). Gels were western blotted and detected with native (N) and denatured (D) 

aldolase B antisera. Since antisera for denatured aldolase B were most sensitive and 

specific for ubiquitinated aldolase B on western blots (Fig. 3-2a and 3-3), these 

antibodies were used to probe for ubiquitinated aldolase B (Fig. 3-1 la, upper panel). A 

major stable ubiquitin-aldolase B conjugate occurred at 68 kD consistent with Ub68 

found in rat liver (Fig. 1-3 and 3-3) and in Fao hepatoma cells (Fig. 1-4). This confirmed 

that E36AB and ts20AB cells contained ubiquitinated aldolase B, including Ub68. 

Similar levels of Ub68 were detected in all conditions, including heat stressed ts20 cells 

(ts20AB, HS MEM) in which ubiquitination is inhibited. Unchanged ubiquitinated 

protein under conditions of inhibited ubiquitination appears contradictory. Though 

greatly inhibited, a low level of ubiquitination continues in heat stressed ts20 cells, but 

this low level is insufficient to mediate cellular processes (Hischberg and Marcus, 1982; 

Kulka, et al., 1988; Gropper, et al., 1991). The results here suggest that low levels of 

ubiquitination were sufficient to maintain multiubiquitinated intermediates of aldolase B 

Figure 3-11: Ubiquitinated Aldolase B Occurs in E36 and ts20 Cells Expressing RABM . 
Replicate sets of E36AB and ts20AB cultures were treated with indicated media 
(DMEM +FBS or MEM) and temperatures (CT or HS), harvested, and 
immunoprecipitated with 9E10 antibody; the immunoprecipitate was pelleted (P) from 
the lysate and proteins remaining in the supernatant were precipitated with 
trichloroacetic acid (S); P and S samples were boiled in 2 x sample buffer, split in equal 
aliquots, and duplicate gels run on SDS-PAGE; a) Western blots to detect RABM 
immunoreactivity (Aldolase B) were made from one gel and upper and lower portions of 
blots were immunodetected with anti-denatured (upper panel, D) and anti-native (lower 
panel, N) aldolase B, respectively; b) the duplicate gel was Coomassie stained, lane 
numbers at the top of the Coomassie gel correspond to identical samples in lanes 
numbered at the bottom of the blots in (a); MW, molecular weight markers; molecular 
weights (kD) are indicated at right. 








P S P S 

P S P S P S 

■■■■ ■■■ ■■■■■ ■■■■■■■■ B^ H tm BH ■ 

P S 






<< < 


1 2 3 4 5 6 7 8 9 10 11 12 

b. MW 1 2 3 4 5 6 7 8 9 10 11 12 MW 

-jjP— - -J 

*~ - ■ 

:;<^B^B^M: <>>^H 

■v Is 



.2 ► 



1 ► 
§ ► 



' tB» 



1 1 1 I 

4 66.2 

^ 55.0 


< 31.0 

^ 21.5 

(Ub68 probably contains 4 ubiquitins), but were insufficient to mediate the proteolysis of 

RABM detected in HP fractions (Fig. 3-10b, lane 7). The occurrence of Ub68 in 

different samples from different cell types (Figs. 1-3, 1-4, and 3-11) supports it as a 

stable basal intermediate that probably requires more ubiquitination to facilitate 

proteolysis. Alternatively, heat stress-induced ubiquitination could operate on the 

machinery of stress-induced degradation. Further experiments are needed to distinguish 

these alternatives. 

As detected with antiserum to native aldolase B (N), immunoprecipitation with 

the 9E10 epitope (specific for RABM) effectively pelleted all detectable 41.3 kD RABM 

protein and more than half of a dim-labeled 40 kD protein, presumably endogenous 

aldolase A (Fig. 3-1 la, lower panel). The 40 kD aldolase A was shown to lack 9E10 

immunoreactivity (Fig. 3-9), indicating that RABM and endogenous aldolase A occur as 

a complex in E36AB and ts20AB cells. This is consistent with the known tetrameric 

structure of all FDP aldolase isozymes, wherein different subunits randomly and stably 

oligomerize during synthesis (reviewed in Chapter 1). Ubiquitinated forms of aldolase B 

were also removed from lysates by 9E10 immunoprecipitation (Fig. 3-1 la, upper panel), 

suggesting that ubiquitinated RABM retained its C-terminal epitope tag or retained 

associations with unmodified 9E10-immuno-reactive RABM subunits. If ubiquitinated 

RABM is not associated with other aldolase subunits, then this would provide evidence 

that ubiquitination might disassemble quaternary structure of aldolase B perhaps as an 

early step in the degradative pathway. However, this possibility was not pursued here. 






Heat Stress 


Heat Stress 


Heat Stress 
ts20 mutant 



Early Lysosomal Late 

Intermediates Proteolysis Intermediates 

basal Aldolase A 

(trace, 41.3 kDRABM) 

^Aldolase A 
(trace, 41.3 kDRABM) 

U Aldolase A 
Lf 41.3 kDRABM 

basal Aldolase A 

40 kDRABM 

40 kDRABM 

40 kD RABM 

40 kDRABM 

41.3 kDRABM 

Figure 3-12: Summary of Association and Limited Proteolysis of RABM and 
Endogenous Aldolase A in Pelleted Organelles of E36 Cells . Each pathway corresponds 
to experimental treatment described in boxes at left and correspond to the following 
abbreviations used above: A) E36AB or ts20AB, CT; B) E36AB, HS; C) E36AB, HS 
+ CHQ; D) ts20AB, HS. Since aldolase A was detected by enzymatic activity and is 
inactivated by limited proteolysis, no late intermediates of Aldolase A are shown. Since 
RABM detected on western blots shifts from 41.3 kD to 40 kD forms by limited 
proteolysis, these forms are listed for early and late intermediates, respectively. Weight 
of white vertical arrows indicate relative increases in detected levels of aldolase A and 
RABM pelleted with organelles. Weight of black horizontal arrows indicate relative 
rates proposed for processes listed in the heading. 

Figure 3-12 summarizes the results of subcellular fractionation studies for 
endogenous aldolase A and RABM expressed in E36 cells. Aldolase A activity is very 
sensitive to proteolytic inactivation and loses 98% of its activity upon limited proteolysis 
(Penhoet and Rutter, 1975; Horecker, et al., 1985). Since this made proteolyzed 

aldolase A undetectable in the background of active aldolase A from E36 cells, aldolase 

A activity was used to demonstrate association with organelles but not for detecting 

proteolyzed intermediates of degradation. RABM was detected by western blotting with 

antiserum to native aldolase B, allowing detection of limited proteolysis (41.3 kD — > 40 

kD) products referred to here as "late intermediates" (Fig. 3-12, last column). Smaller 

molecular weight intermediates of proteolysis were not detected, because they are not 

recognized by anti-native aldolase B and probably are more rapidly degraded than 40 kD 

aldolase B (Reznick, et al., 1985; Horecker, et al., 1985). 

Under control conditions (Fig. 3-12, A), basal levels of aldolase A, 40 kD 
RABM, and a trace of 41 .3 kD RABM were detected in pelletable organelles. Heat 
stress (Fig. 3-12, B) caused a partial increase in aldolase A activity with little apparent 
change in RABM forms. However, a partial increase in trace levels of 41 .3 kD RABM 
were likely to be missed, because they were below the threshold for optimal Enhanced 
Chemiluminescent detection (Amersham). Consistent with reaching the threshold for 
detection, chloroquine inhibition of limited proteolysis caused a sudden signal increase in 
41.3 kD RABM (Fig. 3-12, C). Chloroquine also caused an even more aldolase A to 
accumulate. The results indicated that lysosomal proteolysis mediates loss of aldolase A 
and aldolase B (RABM) associated with organelles, demonstrating sequestration of these 
proteins into lysosomes. 

When ubiquitination was inhibited by the ts20 mutation (Fig. 3-12, D) different 
results were obtained for aldolase A and aldolase B (RABM). Consistent with 

ubiquitination mediating degradative mechanisms after sequestration (Lenk, et al., 1992), 

41.3 kD RABM accumulated in lysosomes. Under the same conditions, aldolase A 
activity did not accumulate. For this to happen, reduced ubiquitination in ts20 cells 
would have to increase lysosomal proteolysis or decrease sequestration of aldolase A 
into lysosomes. Since the former possibility is unlikely, aldolase A probably requires 
ubiquitination for sequestration as well as proteolysis. Since assays for aldolase A and 
RABM were fundamentally different, the difference in ubiquitination effects between the 
two isoforms might be due to differences in detection by enzymatic activity versus 
immunoreactivity. Further experiments are needed to address this problem, but the 
results here suggest that both aldolase A and aldolase B require ubiquitination for 
degradation in lysosomes. 

Ubiquitin-Mediated Autophagic Degradation Occurs in E36AB Cells 
Above, evidence supporting ubiquitin-mediated proteolysis of aldolase B 
(RABM) in lysosomes during heat stress was presented. To confirm that ubiquitin- 
mediated autophagic degradation occurred during the fractionation studies above, 
degradation rates of total long-lived proteins were measured for E36AB and ts20AB 
cells under the same conditions. All protein degradation rates reported in this 
dissertation are calculated as for degradative rate constants. Procedures are described in 
Figure 3-13 and Materials and Methods. Arbitrary examples of regression analysis for 
basal (control) and induced (stress) degradation rates are indicated in Figure 3-13. The 
degradative rate constant, ka, is defined as the negative slope (coefficient of x) of the 

Figure 3-13: Quantification of Degradation Rates , a) Cells were metabolically 
radiolabeled and chased as described in Materials and Methods. Radiolabel signal (less 
background) in TCA precipitates (for total long-lived proteins) or at appropriate 
molecular weights for immunoprecipitates run on SDS-PAGE (for aldolase B) was 
measured and magnitudes were natural-log transformed and plotted as described in 
Materials and Methods. Regression lines, equations, and R 2 correlation statistics are 
shown for arbitrary examples of basal (Control) and heat stressed (Stress) protein 
degradation, b) Slopes reported as decimal numbers in (a) are plotted as bars and 
reported with units of %/h (= kd (/h) x 100%). The SEM of the slope was calculated 
using Microsoft Excel Spreadsheet Analysis package: ANOVA regression error 
(standard error of y- values at given x- values divided by the square root of the deviations 
squared of x-values). 




















, i 

y = -0.01 12x + 4.5339 
R 2 = 0.9859 

y = -0.0322x + 4.5737 
R 2 = 0.9061 


5 10 15 

Chase Time, hours 

a Control • Stress 
Linear (Control) Linear (Stress) 


b. □ control m stress 

— \s.\j /O 


~ 3.0% - 



£ 2.5% - 



§1 2.0% - 


2 1.5% - 


2 1.0% - 



» 0.5% - 



"• n n% 

u.u/o ^ 

Time Course as a Bar, 

Negative Slope 

+/- Slope Error 

lines plotted in Figure 3- 13a and of signal decrease in this dissertation are calculated as 

done for kj. Note of caution: different methods of calculating degradation are used by 

different investigators which would give different rate magnitudes for the same data; 

rates presented here are based on the empirically supported assumption that protein 

degradation mechanisms operate with first order kinetics. 

In this chapter, degradation rates are presented for total trichloroacetic acid 
(TCA) precipitated proteins. Such rates represent an average of the pooled degradative 
rate constants (kj) for all radiolabeled proteins over the time interval of the assay. 
Variation in metabolic radiolabeling and chase intervals yield differentially labeled total 
proteins due to widely different turnover rates for individual proteins, and such variation 
would change the pooled kd. These factors were controlled to give comparable rates 
between different experiments throughout this dissertation. 

To confirmed that E36 cells undergo heat stress enhanced degradation of total 
protein that requires ubiquitination, E36 cells, ts20 cells, and ts20El cells were 
radiolabeled then chased under control (CT) and heat stress (HS) conditions, and 
degradation rates were calculated as described above (Fig. 3-13). E36 cells were found 
to undergo about a two fold enhancement in the degradation of long-lived proteins (Fig. 
3-14, E36). The ts20 cell line derived from E36 cells has a temperature-sensitive 
mutation in El ubiquitin-activating eruiyme, and the heat stress (HS) is non-permissive 
for El function in ts20 cells, greatly inhibiting ubiquitination (Kulka, et al., 1988) and 
ubiquitin-mediated protein degradation (Gropper, et al., 1991). Consistent with this, the 

reported as the fractional degradative rate in Figure 3- 13b. For consistency and to allow 

direct comparisons of data, all rates HS-enhanced degradation of long-lived proteins was 

blocked in ts20 cells (Fig. 3-14, ts20). Furthermore, the mutant degradative phenotype 

was rescued in ts20 cells permanently expressing wild type human El (Fig. 3-14, 



.2 3% 






= 2% 





^ 1% 


D CT (permissive) 
II HS (non-permissive) 




Cell Line 


Figure 3-14: Heat Stress-Induced Degradation of Long-lived Proteins in E36 Cells 
Requires El -Mediated Ubiquitination . Cultures of parent (E36) cells, El-ubiquitination 
mutant (ts20) cells, and mutant rescued with cDNA for wildtype human El (ts20El) 
cells were assayed for degradation of total TCA precipitated long-lived proteins 
radiolabeled with 14 C-valine (Fig. 3-13, Materials and Methods) reported for the 
indicated conditions (mean ± SEM, n = 9); CT, control temperature; HS, heat stress; *, 
HS > CT (Student's t-test, p < 0.000001). 

The results demonstrate that heat stress-induced degradation in E36 cells is dependent 

on ubiquitination mediated by El . The degradation of long-lived proteins in E36AB and 

ts20AB cells was assayed (Fig. 3-15), and the results were similar to those for 

untransfected E36 and ts20 cells (Fig. 3-14). This demonstrated that RABM expression 

had no effect on the degradative phenotypes in these cells. Cultures were also treated 

with the relatively specific inhibitor of autophagy, 3-methyladenine (3 MA), or with the 

lysosomal hydrolase inhibitor, chloroquine (CHQ). Both inhibitors effectively blocked 

the enhanced degradation of long-lived proteins (Fig. 3-15). The results confirmed that 

heat stress-induced degradation of most long-lived proteins occurred in lysosomes via 


In Figure 3-15, ts20AB cells appear to have reduced basal degradation relative 

to E36AB cells. This result was not consistent between experiments, and in general, 

E36 and ts20 cells had similar basal degradation rates that were unaffected by RABM 

expression (data not shown). Chloroquine (CHQ) and 3-methyladenine (3MA) inhibited 

a fraction of basal protein degradation at the concentrations shown (Fig. 3-15). Though 

the effect of chloroquine was marginal here (40 u,M), at higher concentrations (80 & 160 

uM) inhibition of basal degradation was more significant (data not shown). CHQ at the 

higher concentrations and 3MA at the least inhibitory concentration reduced the viability 

of ts20 cells during heat stress (Materials and Methods; data not shown), so degradation 

measurements for such treatments were not included. As indicated by sensitivity to 

CHQ or 3MA, 17±6% of basal degradation of long-lived proteins occurred by lysosomal 

degradation via autophagy, corresponding to a rate of 0.3 ± 0. 1%/h (mean ± SEM, n = 


12). Given the estimated basal rate of autophagy, heat stress (HS) caused a calculated 
6.6 ± 1.0 fold induction over basal autophagic degradation in uninhibited wildtype E36 

E36AB E36AB E36AB ts20AB ts20AB 


Cell Line & Inhibitor 

Figure 3-15: Heat Stress-Induced Autophagic Degradation of Long-lived Proteins that 
Requires Ubiquitination Occurs in Cells Expressing RABM . Cultures of E36AB and 
ts20AB cells were metabolically labeled with 35 S-methionine, and at various times of 
chase, cultures were lysed in modified RIP A buffer and 2% of the lysate was TCA 
precipitated and processed to measure degradation of long-lived proteins (mean ± SEM, 
n = 16-34) as previously described (Fig. 3-13 and Materials and Methods); CT, control 
temperature; HS, heat stress; indicated chase medium were supplemented with 40 uM 
chloroquine (+CHQ) or 5 mM 3-methyladenine (+3MA). Significant differences 
(Student's t-test, p < 0.05): mean < untreated CT in corresponding cells (*, p < 0.05 
•*, p < 0.02; ***, p < 0.008); ****, E36AB, HS > CT (p < 0.0002). The remaining 
lysates (98% of the volume) were used to immunoprecipitate and measure the 
proteolysis of RABM discussed below (Fig. 4-4). 

cells (mean ± SEM, n = 3). In ubiquitination mutant ts20 cells, enhanced degradation 

was blocked (Figs. 3-14 and 3-15; Gropper, et al., 1991), and a 6-fold increase in 

fractional volume of autophagic vacuoles occurred (Lenk, et al., 1992). Together, the 

data support a model in which intralysosomal complete degradation of proteins requires 

ubiquitination, but sequestration does not (Lenk, et al., 1992). 

Basal autophagic degradation explained why aldolase A and RABM were found 

with pelletable organelles in all conditions (Fig. 3-4a and 3-10). Neither aldolase A 

activity nor RABM immunoreactivity accumulated in HS ts20 cells, even though 

autophagic sequestration was increased 6-fold (previous paragraph). Ubiquitination was 

required for limited proteolysis of RABM from 41.3 kD to 40 kD (Fig. 3- 10b), but other 

limited proteolysis appeared sufficient enough to destroy enzyme activity and 

immunoreactivity, preventing multifold accumulation of detectable aldolase. Though 

detectable aldolase fragments did not accumulate, proteolysis did not seem to proceed to 

amino acids, because enhanced lysosomal degradation of TCA-precipitable polypeptides 

was blocked in HS ts20 cells (Figs. 3-14 and 3-15). The data are consistent with 

ubiquitination being required for a subset of lysosomal protease activities needed to 

completely degrade long-lived proteins to amino acids, and one such protease can be 

detected as intralysosomal limited proteolysis of RABM (aldolase B). 




In the literature, starvation-induced autophagic degradation of aldolase B has 
been proposed, and heat stress-induced autophagic degradation of long-lived proteins in 
E36 cells was shown to require ubiquitination. As shown above, ubiquitinated aldolase 
B (Ub68) occurred in cultured E36 cells and in vivo enriched with organelles of 
starvation-induced autophagy isolated from rat liver. In E36 cells, assays for aldolase 
associated with organelles showed that lysosomal accumulation of endogenous aldolase 
A and lysosomal proteolysis of exogenous aldolase B (RABM) required ubiquitination. 
The results are consistent with degradation of aldolase isoforms in general occurring by 
ubiquitin-mediated autophagic degradation inside lysosomes. However, aldolase A and 
aldolase B have been proposed to undergo limited proteolysis on the outside of 
lysosomes independent of acidic pH (Pontremoli, et al., 1982; Horecker, et al., 1985; 
Sygush, et al., 1990). To examine the contribution of cytosolic and lysosomal 
mechanisms, rates for basal and heat stress-induced proteolysis of aldolase B (RABM) in 
E36 cells were quantified in the presence of lysosomal inhibition. The results below 
show that temperature-dependent cytosolic proteolysis of aldolase B was independent of 
ubiquitination, that complete cytosolic protein degradation to amino acids was 


independent of temperature, and that autophagic degradation also increased in a 

temperature-dependent manner. 

Ubiquitin- Independent Cytosolic Proteolysis of Aldolase B 
To measure proteolysis of aldolase B, the decrease of radioactivity in RABM 
protein bands on gels of 9E10-immunoprecipitates isolated from metabolically labeled 
cells during pulse-chase experiments was measured. Immunoprecipitation was efficient 
with anti-9E10 (Fig. 3-1 la, lower panel, N) and consistently yielded more RABM than 
for anti-aldolase B (Fig. 4-1). 9E10 immunoprecipitation gave the most reproducible 
data and was used extensively for experimental repetitions; unless otherwise stated, 
degradation rates for RABM are reported for such data. Standard curves were done to 
confirm a linear response for quantification of the radiolabeled RABM (Fig. 4-2). 
Degradation rates were calculated in the same manner as for total long-lived protein 
degradation using RABM-specific signal (Fig. 3-13). Since a specific protein was 
measured, the resulting degradation rates were equivalent to the degradative rate 
constant (kd) for RABM (aldolase B) proteolysis. The results below agree well with 
published kd for aldolase B degradation in cultured cells (1.0-2.5%/h basal and 3.0- 
5.0%/h stress-induced). 

To determine the contribution of ubiquitination in heat stress-induced 
degradation of aldolase B, degradation rates for RABM in E36AB and ts20AB cells at 
control(CT) and heat stress (HS) temperatures were measured (Fig. 4-3). The 



Aldolase B 
E36AB ts20AB ts20AB ts20 


Figure 4-1 : Immunoprecipitation of RABM With Antibodies Against 9E10 or Native 
Aldolase B Epitopes . Anti-myc-sepharose in mRIPA buffer lysates (9E10) or aldolase B 
antiserum precipitated with protein A-agarose in SLB lysates (Aldolase B) were used to 
immunoprecipitate radiolabeled RABM from E36AB and ts20AB cell lines; ts20 cells 
lacking RABM demonstrate background for the least stringent immunoprecipitation 
(Aldolase B). Arrowheads indicate only specifically precipitated bands (40 and 41.3 
kD), migrating in a region of consistently low background. The prominent 44-45 kD 
band in 9E10 immunoprecipitations was non-specific, representing actin that has a high 
affinity for sepharose (Sigma technical support). 




































y= 0.991 8x+ 0.0318 
R 2 = 0.9287 



0% 20% 40% 60% 80% 100% 

Fraction of 9E10-lmmunoprecipitation Lysate 
from RABM Expressing Cells 

Figure 4-2: Standard Curve For 9E10 Immunoprecipitation of RABM . Confluent 
cultures of ts20AB and ts20 (untransfected) cells were metabolically labeled with 35 S- 
methionine, lysed in mRIPA buffer, lysates precleared, and immunoprecipitated (anti- 
9E10 conjugated to sepharose) done as described in Materials and Methods, except 
ts20AB lysate (containing maximal RABM) was diluted with ts20 lysate (lacking 
RABM) as indicated. Each immunoprecipitation was done from -400,000 cell 
equivalents of lysate. Regression line and equation for least squared deviation is shown 
with R 2 correlation statistic; R = 0.9637 (ideal would be a slope of 1.00 with intercept of 
0.00andR 2 =1.00). 

degradation of RABM was two fold enhanced by heat stress. Though starvation- 
induced degradation of aldolase B is established (reviewed in Chapter 1), this is the first 
demonstration that aldolase B undergoes enhanced degradation during heat stress. 
Furthermore, induced proteolysis continued even when ubiquitination was inhibited (Fig. 
4-3 HS,ts20AB), indicating involvement a ubiquitin- independent mechanism. This 
suggested that in addition to ubiquitin-mediated lysosomal proteolysis (Fig. 3-12), a 

second proteolytic mechanism for RABM occurred at rates similar to heat stress-induced 

autophagy that required ubiquitination (Fig. 3-15). 



Cell Line 

Figure 4-3: Heat Stress-Induced Degradation of RABM Occurs Independently of El- 
Mediated Ubiquitination . Indicated cell lines were metabolically labeled with 35 S- 
methionine, chased for various times, immunoprecipitated, radioactivity at the molecular 
weight of RABM was quantified, and regression analysis performed to calculate the 
degradative rate constant ± standard error (n = 3 1 to 34) for RABM as described in 
Materials and Methods, under control (CT) and heat stress (HS) conditions. Reminder: 
HS is non-permissive for El -mediated ubiquitination in ts20AB cells. For both cell lines, 
RABM degradation at HS was higher than at CT (p < 0.0007). 

Temperature-Dependent Cytosolic Proteolysis in Fao Cells 

To determine how much heat stress-enhanced proteolysis of RABM occurred in 

lysosomes, chloroquine (CHQ) was used to inhibit lysosomal hydrolases (Fig. 4-4). 

0% H 






Cell Line + Drug 

Figure 4-4: Heat Stress-Induced Degradation of RABM Occurs Independently of 
Lysosomal Hydrolasas and Autophagy . Cultures were treated and processed as for 
Figure 4-18 to calculate degradation rates (mean ± SEM, n = 16 to 34), except cultures 
were treated with either 5 mM 3-methyladenine (+3MA) or 40 u.M chloroquine (+CHQ) 
to inhibit autophagy or lysosomal hydrolases, respectively. Student t-tests: for all 
treatments, HS > CT (p < 0.0013); *, 3MA partially inhibited HS degradation (p < 
0.03 1); all other differences were not significant at a 0.05 level. Note: 2% of each 
culture lysate was TCA precipitated and used to measure degradation of long-lived 
proteins occurring under identical conditions (Fig. 3-15). 

Chloroquine failed to reduce the heat stress-induced degradation rate for RABM; note, 

the apparent increase shown in Fig. 4-4 was neither statistically significant nor 
reproducible (data not shown). Even when ubiquitination and lysosomal degradation 
were simultaneously inhibited (ts20AB+CHQ), enhanced degradation of RABM still 
continued (Figs. 4-4). This indicated that heat stress-induced proteolysis of aldolase B 
in E36 cells was independent of ubiquitin-mediated autophagy and lysosomal 
degradation. Supporting this, 3-methyadenineonly marginally inhibited heat stress- 
induced autophagic degradation (Fig. 4-4), and the marginal inhibition was accounted by 
a comparable reduction of basal protein degradation observed for samples from the same 
cultures (Fig. 3-15). These results support a role for cytosolic proteolysis in the heat 
stress-induced degradation of aldolase B. 

The inhibitor resistant proteolysis of RABM (Fig. 4-4) seemed to contradict the 
inhibitor sensitive degradation of total long-lived proteins measured for the same cultures 
(Fig. 3-15), suggesting that degradation of RABM occurred by a selective cytosolic 
mechanism which was specific for a fraction of total long-lived proteins that includes 
aldolase B. An alternative explanation is possible, because protein degradation 
measurements using TCA and immunoprecipitation are fundamentally different from 
each other. Degradation measurements for total proteins utilize TCA precipitation 
which cannot distinguish full length protein from polypeptide fragments created by 
partial proteolysis. However, immunoprecipitation is sensitive to limited proteolysis, 
because a single proteolytic cleavage can disrupt the size and structure of a protein 

removing it from an SDS-PAGE band used to quantify degradation. Furthermore, 

aldolase B is known to undergo proteolysis to relatively large polypeptide fragments that 

cannot be recognized by antibodies against the native protein (Reznick, et al., 1985), 

demonstrating that partial proteolysis can disrupt immunoreactivity. Immunoreactivity 

lost to partial proteolysis causes loss of radiolabel from corresponding 

immunoprecipitates, including fragments that might be quite large. Thus, 

immunoprecipitation effectively detects limited proteolysis whereas TCA precipitation 

requires complete degradation to amino acids. Given this, cytosolic proteolysis of 

aldolase B (RABM) occurred independently by partial proteolysis that produced 

undetectable polypeptide fragments (TCA precipitable) requiring autophagic degradation 

in lysosomes for complete proteolysis. 

Since polypeptide fragments of aldolase B could not be detected the possibility 

remains that aldolase B underwent an enhanced complete degradation in cytosol during 

heat stress. However, this would have to be a unique pathway for aldolase B because 

for most proteins (TCA precipitated) autophagic degradation was required. Given this, a 

cytosolic mechanism for complete degradation of aldolase B would have to be selective 

for aldolase B, requiring s recognition mechanism that is novel and unsubstantiated in the 

literature. This seem less likely than a more general model in which many proteins, 

including aldolase B, undergo partial cytosolic proteolysis followed by autophagic 

degradation in lysosomes. 

Degradation assays using anti-aldolase B immunoprecipitation gave similar 

results to anti-9E10 (data not shown). This indicated that proteolysis of RABM was not 

limited to the epitope tag but included at least some degradation of aldolase B protein. 

Since immunoprecipitations were done from extracts of cell sheets, loss of RABM to the 

media would look like degradation. To control for this, heat stress release of aldolase 

activity into media was measured and found statistically insignificant compared to 

variation in induced degradation measurements (Fig. 4-5). 



«,- 25% 

i 20% 


■o 1.9% 



io 1.0% 

2 0.5% 



1 0.0% 

— 1 1 1 





Release to 


Measirerrent of Aldolase 

Figure 4-5: Heat Stress-Induced Degradation of Aldolase B was Not Due to Release of 
Aldolase to Medium . During chase treatments, aliquots of media were assayed for 
aldolase activity; after the final chase, cells were harvested, total activity in the culture 
was calculated, and the fractional rate of activity loss calculated. The rate calculation 
and measurements for the degradation of aldolase B are described in Materials and 
Methods. Release of aldolase to media > (p < 0.034) but had a low probability of 
contributing an effect on degradation rates (Student's t-test, p < 0.0001). At control 
temperatures (data not shown), aldolase activity released to the medium was not 
significantly greater than zero (p > 0.4). 

Starvation-Induced Autophagic Degradation of Aldolase B in Fao Cells 
To control for the nature of the epitope tag on RABM and determine whether 
cytosolic proteolysis of aldolase B generally occurred in different cells, heat stress- 
induced proteolysis of epitope-tagged human aldolase B (HAHAB) expressed in Fao 
cells was examined (next section). HAHAB carried a 12CA5 HA (hemagglutinin) 
epitope on its amino terminus. Measuring degradation of HAHAB in Fao cells versus 
RABM in E36 cells controlled for cell type (rat hepatoma vs. Chinese hamster lung), 
position and identity of the epitope tag (amino-HA vs. carboxyl-myc), and the source 
species of the recombinant aldolase B (human vs. rat). Whether heat stress can induce 
autophagy in Fao cells was not known. Before examining this possibility, it was 
reasonable to confirm that starvation-induced autophagic degradation of exogenous 
aldolase B (HAHAB) occurred like that for endogenous aldolase B. 

Fao cells retain many characteristics of differentiated hepatocytes, including 
expression of endogenous aldolase B (Deschatrette et al., 1979; Deschatrette and Weiss, 
1974), and starvation-induced degradation of long-lived proteins (Fig. 4-6, bar 2). 
Treatment of Fao cultures with 3-methyladnine (3MA) to block autophagic 
sequestration, chloroquine (CHQ) to inhibit lysosomal acid hydrolases, or leupeptin 
(LEUP) to inhibit lysosomal serine proteinases blocked starvation-induced degradation 
(Fig. 4-6), demonstrating that enhanced degradation of long-lived proteins occurred in 
lysosomes via autophagy in starved Fao cells as well as in heat stressed E36 cells. E36 
cells did not undergo starvation-induced degradation (data not shown), but Fao cells 


provided an opportunity to examine both starvation-induced and heat stress-induced 
degradation of aldolase B in the same cell type. 

Fed Starved Starved Starved Starved 


Figure 4-6: Starvation Induces Degradation of Long-lived Proteins in Fao Cells . Fao cell 
cultures were metabolically labeled with 14 C-valine, chased with DMEM + 10% FBS 
(Fed, bar 1) or Krebs-Heinseleit Medium (Starved, bars 2-5) containing "cold" 10 mM 
valine, and loss of TCA precipitable counts was used to calculate degradation rates 
(mean ± SEM, n = 8-16) as described in Materials and Methods; Krebs-Heinseleit is an 
amino acid free minimal medium; 3MA (bar 3), 10 mM 3-methyladenine; LEUP (bar 4), 
300 mM leupeptin; CHQ (bar 5), 160 uM chloroquine. Significant differences 
(Student's t-test, p < 0.0001): bar 2 > bars 1, 3, 4, or 5; bar 5 < bar 1. 




Figure 4-7: Permanent Expression of HAHAB in Fao Cells . Fao cells were transfected 
and selected for permanent expression of human aldolase B carrying a 12CA5 HA 
epitope tag on its amino terminus (HAHAB), resulting a cell line permanently expressing 
HAHAB (FaoAB). Untransfected Fao (Fao) and FaoAB cells were processed for 
12CA5 immunofluorescence microscopy (upper panels) as described in Materials and 
Methods; Phase (lower panels), phase contrast of same field as upper panels. Scale bar = 
50 uM. 

To effectively measure the degradation of aldolase B in Fao cells, an Fao cell line 

permanently expressing epitope-tagged aldolase B (HAHAB) was isolated. Figure 4-7 

shows the immunofluorescent signal for the HA epitope (12CA5) in a cell line 

permanently expressing HAHAB (FaoAB). Note the bright signal in FaoAB cells 

relative to the untransfected parent cells (Fao). Anti-HA epitope (12CA5) specifically 

detected expression of tagged exogenous aldolase B as indicated by lack of label in 

untransfected cells (Fig. 4-7, Fao). The presence of negatively labeled nuclei and 

vacuoles supported a cytosolic localization for HAHAB. Consistent with Fao cells 

expressing endogenous aldolase B, Fao and FaoAB cells stained with anti-aldolase B had 

similar labeling patterns to each other (data not shown) which resembled the transfected 

cells labeled with anti-12CA5 (Fig. 4-7, FaoAB). 

Expression of HAHAB was confirmed on western blots (Fig. 4-8). Western blot 

analysis with anti-aldolase B (Fig. 4-8a) showed 40 kD endogenous aldolase B in 

untransfected Fao cells (Fao). FaoAB cells (FaoAB) also contained a 41.3 kD aldolase 

B-immunoreactive protein indicating HAHAB expression. Non-specific background and 

a 41 .3 kD marker for epitope-tagged aldolase B were provided by ts20 and ts20AB cells 

that lack endogenous aldolase B. HAHAB comigrated with RABM, consistent with 

12CA5 and 9E10 epitope constructs being similar in size. Fao cells express over 10 fold 

more aldolase A than aldolase B (Deschatrette et al., 1979; Deschatrette and Weiss, 

1974), so the pattern of labeling was consistent with HAHAB expression equal to or less 

than total endogenous aldolase (isoforms A + B). 



FaoAB Fao 


Aldolase B 

- Western 



























Figure 4-8: Biochemical Detection of HAHAB . a) anti-aldolase B western blot analysis 
was performed on whole cell extracts from the indicated cell lines as described in 
Materials and Methods; ~1 cm 2 culture area equivalent loaded per lane from different 
confluency: ts20AB and ts20 lanes from tightly confluent, first FaoAB and last Fao lanes 
from recently confluent, remaining lanes from subconfluent; ts20 cells have no 
endogenous aldolase B, and ts20AB cells express a different epitope-tagged aldolase B, 
RABM; b) autoradiograph of duplicate anti-12CA5 (HA epitope) immunoprecipitates 

prepared from S-methionine labeled FaoAB and Fao cultures as described in Materials 
and Methods. Molecular weights for proteins specifically detected by the methods are 

On autoradiographs of anti-12CA5 immunoprecipitates (12CA5-IP) from cells 

metabolically labeled with radioactive amino acid, a doublet of HAHAB (41.3 kD) and 

endogenous aldolase (40.0 kD) was specifically precipitated from FaoAB cells and not 

Fao cells (4-8b). Since endogenous aldolase lacked 12CA5 immunoreactivity (Fig. 4-7, 

Fao), the 40 kD aldolase subunits appeared to be associated with HAHAB for 

immunoprecipitation. This provided evidence that HAHAB subunits tetramerized with 

endogenous subunits. The results indicated that HAHAB was permanently expressed in 

a manner expected for a subunit of aldolase. Furthermore, aldolase protein bands 

migrated in a region of low background on SDS-PAGE of immunoprecipitates 

facilitating quantification of radioactivity associated with HAHAB (Fig. 4-8b). 

In order to compare stress-induced degradation of HAHAB in Fao cells to 

previous results for RABM in E36 cells (Fig. 4-3 and 4-4), 12CA5-immunoprecipitation 

was used to isolate and measure the degradation of HAHAB (Fig. 4-9). Starvation 

increased the degradation of HAHAB by at least two fold (Figs. 4-9 and 4-11) which 

was even greater than the enhancement of total long-lived protein degradation (Fig. 4- 

6). Previously, heat stress-induced degradation of RABM (aldolase B) in E36 cells 

resisted lysosomal inhibitors (Fig. 4-4), but starvation enhanced degradation of HAHAB 

was completely inhibited by chloroquine (CHQ) or 3-methyladenine (3 MA) in Fao cells 

(Fig. 4-9). The data indicated that HAHAB was degraded in lysosomes via autophagy 

during starvation, and that cytosolic proteolysis was constant at the constant temperature 

(37°C) during starvation in E36 cells. 












4% - 





■ Starved 

H 1 

x Egg 





Inhibitor Treatment 


Figure 4-9: Starvation-Induced Autophagic Degradation of HAHAB in Fao Cells . 

FaoAB cells labeled with S-methionine were chased with excess "cold" methionine in 
DMEM + 10% FBS (Fed) or Krebs-Heinseleit medium without amino acids (Starved) 
with 160 uM chloroquine (+CHQ), 10 mM 3-methyladenine (+3MA), or no additions 
(None); radiolabel in immunopurifed HAHAB was used to calculate degradation rate 
(mean ± SEM, n = 1 1-23) as described in Materials and Methods; *, HS > CT, 
Student's t-test (p < 0.0005). 

To confirm that HAHAB was representative of endogenous aldolase B, 
starvation-induced loss of endogenous aldolase B was examined on western blots (Fig. 
4-10). When protein synthesis is inhibited protein decreases reflect degradation rates 
(Henell, et al., 1987). In fed confluent cultures (Fed), aldolase B levels remained 
constant or slightly increased (statistically insignificant) during 8 hour incubations (Fig. 

4-10). Starvation inhibited protein synthesis (data not shown), and endogenous aldolase 

B levels decreased at a rate that was statistically indistinguishable from the enhanced 

degradation of HAHAB (Fig. 4-9 and 4-10). 3MA or CHQ inhibited the loss of 

endogenous aldolase B to a rate of ~2%/h (Fig. 4-10), consistent with basal cytosolic 

degradation (Fig. 4-9) and indicating that endogenous aldolase B and HAHAB utilized 

same starvation-induced autophagic mechanism for degradation. 




+3 MA 


Figure 4-10: Starvation-Induced Autophagic Degradation of Endogenous Rat Aldolase 
B in Fao Cells . Untransfected Fao cells were incubated in the chase media of Figure 4-9 
(same labels) for various times, and whole cells directly harvested by boiling in sample 
loading buffer were analyzed on anti-aldolase B western blots, and changes (mean ± 
SEM, n = 16-24) in the level of aldolase B were calculated as previously described 
(Materials and Methods). Rates of decrease were plotted, so negative value indicates 
aldolase B increase. *Starved > Fed (Student's t-test, p < 0.0001); ** 
Starved+Inhibitor < Starved (p < 0.0007); ***Fed = 0%/h (p > 0.15). 


Temperature-Dependent Autophagy and Cytosolic Proteolysis 

Due to the temperature-sensitive mutation in ts20 cells, E36 and ts20 cells were 
routinely maintained at 30.5°C, according to established protocols used here (Gropper, 
et al., 1991). For the established experimental system, control temperatures were 30.5°C 
(CT) and heat stress was 41.5°C for one hour then 39.5°C (HS). Heat stress and 
starvation at 37°C, both, enhanced the complete degradation of total TCA precipitated 
polypeptides by autophagy (Figs. 3-15 and 4-6). In agreement with this, starvation 
enhanced the autophagic degradation of immunodetected aldolase B at 37°C (Figs. 4-9 
and 4-10). These results indicated that starvation-induced autophagic degradation of 
aldolase B was rate limiting and faster than cytosolic proteolysis of aldolase B at 37°C in 
Fao cells. In E36 cells, heat stress increased cytosolic proteolysis of immunoprecipitable 
aldolase B to a rate greater than heat stress-induced autophagic degradation (Figs. 3-14, 
3-15, 4-3, and 4-4). The cytosolic proteolysis, though probably partial, disrupted 
immunoprecipitation of aldolase B such that degradation of aldolase B appeared to be 
independent of lysosomal degradation. 

To determine whether heat stress-induced cytosolic proteolysis of aldolase B was 
a general phenomenon beyond E36 cells, the heat stress and starvation protocols were 
compared using Fao cells (Fig. 4-1 1). Consistent with E36 cells, 39.5°C heat (stress) 
caused increased proteolysis of HAHAB relative to 30.5°C control temperature 
(unstressed) in Fao cells (Fig. 4-11, Heat Stress Induction, untreated). The proteolysis 


enhanced by 39.5°C was resistant to chloroquine treatment (CHQ), indicating the 

presence of the enhanced cytosolic mechanism (Fig. 4-11, Heat Stress Induction). Sister 

untreated +CHQ 

Heat Stress Induction 

untreated +CHQ 

Starvation Induction 

Experimental System & Chloroquine (+CHQ) 

Figure 4-11: Temperature Increases Cytosolic Proteolysis and Starvation Induces 
Lysosomal Proteolysis of Aldolase B (HAHAB I Heat stress and starvation induction of 
protein degradation in Fao cells and degradation of HAHAB was done as previously 
described above (mean ± SEM, n = 14-16). "Control" and "Stress" treatments are 
different for the two induction systems. The degradation rates are estimates of Iq for 
HAHAB proteolysis. Student's t-test: *, Stress > Control (p < 0.001). Arrows connect 
temperatures that indicate ka estimates for "cytosolic" (CHQ-resistant) proteolysis. 

cultures of the Fao cells confirmed that starvation-induced autophagic degradation of 
aldolase B occurred at 37°C when amino acids and additional nutrients were withheld 
(Fig. 4-11, Starvation Induction). For the starvation-induced degradation at 37°C, it 

was confirmed that inducible lysosomal degradation was inhibited by chloroquine 

treatment (Fig. 4-11, Starvation Induction, ±CHQ). Altogether, the data demonstrated 

that cytosolic proteolysis of aldolase B can occur in different cell types (E36 Chinese 

hamster lung and Fao rat hepatoma) with different aldolase B constructs (RABM and 

HAHAB), and is temperature-dependent. 

The degradation measurements for aldolase B represent rate constants (k<i) set by 
rate-limiting mechanisms. For cytosolic proteolysis of HAHAB, the rate limiting step 
increased proportionally with increased temperature (Fig. 4-11, arrows and 
temperatures). This suggested that increased reactivity was by direct thermal stimulation 
of a rate-limiting cytosolic mechanism. To test this possibility, the k<j's that were 
determined for three different temperatures (30.5°C, 37°C, and 39.5°C) were used to 
make Arrhenius plots (Fig. 4-12) to test this possibility. 

Evidence derived from the Arrhenius plots (Fig. 4-12) is limited by the fact that 
only three temperatures were available. Total cellular degradation of long-lived proteins 
in cultured smooth muscle cells was previously shown to undergo direct thermal 
stimulation with E a = 18 kcal/mole at 15-37°C (Bates, et al., 1982). With similar data 
here (Total TCA), an E a = 19.6 ± 1.5 kcal/mole at 30.5-39.5°C was calculated and was 
statistically indistinguishable (p > 0.1) from the previously published result (Fig. 4-12). 
This argued that our calculations gave reasonable estimates of E a . 



0.0120 - 

/ r = 0.99.-0 
/ ** ° 

-R), kcal/mol-K 

o o 
o o 

8 5 

° - "" 

/ ^-**V-38.5±4.9 
J..-** X r = 0.93 

S^f r = 0.66 A 



^ / & 

X 0.0090 - 

_ A 

>* A 3 


^ 0.0080 - 

6°' . 

1 " 19.6 ±1.5 


r = 0.98 

■ j ^^ 000 


0.0060 - 

1 1 — 

-1 1 1 — 

H 1 1 1 1 


















^ - 




























































Reciprocal Temperature, 1/K 

Figure 4-12: Heat Stress Induces Cytosolic Proteolysis of Aldolase B By Thermal 
Stimulation of Lysosomes . Cytosolic (Cyto, light unbroken line) or lysosomal (Lyso, 
broken line) proteolysis was degradation resistant or sensitive, respectively, to inhibition 
with 3-methyladenine or chloroquine; the sum of these (Cyto+Lyso) represents total 
cellular proteolysis (Total, heavy unbroken line); degradation rates for immunopreci- 
pitated aldolase B (HAHAB, filled symbols) or TCA precipitated polypeptides (TCA 
open symbols) were used as k d -estimates to construct Arrhenius plots with slope = 
activation energy in kcal/mole (E a ), an arc connects estimated slope, E a (mean ± SEM, n 
= 7 or 8) accompanied by the corresponding correlation coefficient (r) for a least squares 
linear plot; square represents cytosolic proteolysis of HAHAB; triangle, circle, or 
diamond represents cytosolic, lysosomal, or total proteolysis of long-lived proteins, 
respectively. By Student's t-test, all slopes were much > and different from each other 
(p < 0.0004), except "Cyto TCA" was close but slightly > (p < 0.028), and "Cyto 
HAHAB" slope was indistinguishable from "Lyso TCA" (p > 0.53). 

Based on sensitivity to inhibitors, lysosomal (Lyso) and cytosolic (Cyto) degradative 

mechanisms for long-lived proteins were separated (Fig. 4-12, TCA). Both had linear 

Arrhenius plots, giving E a = 40.6 ± 1 .8 and 5.1 ± 2.2 kcal/mole for lysosomal and 

cytosolic degradation, respectively. A plot for the cytosolic proteolysis of HAHAB 

(aldolase B) was linear confirming that heat stress induced degradation could be 

accounted for by thermodynamic effects of temperature. The cytosolic proteolysis of 

HAHAB had an E a = 38.5 ± 4.9 kcal/mole which was much larger than for cytosolic 

proteolysis of TCA precipitable polypeptides but closely corresponded to the E a of a 

lysosomal mechanism reported above, as well as the lysosomal degradation of 

endocytosed apolipoprotein B, E a = 35 kcal/mole at 20-3 7°C (Bates, et al., 1982). 

Together results suggested that "cytosolic" (3MA or chloroquine resistant) proteolysis 

of aldolase B (HAHAB) was independent of autophagic sequestration (inhibition) and 

acidic lysosomal pH but occurred by a complex mechanism with E a similar to lysosomal 

degradative mechanisms. 

Lysosomal degradation of long-lived proteins is generally thought to be limited 

by sequestration rate (high E a ), and sequestration can occur by non-autophagic (possibly 

3MA-resistant) process of receptor recognition and translocation (Dice and Chiang, 

1988; Cuervo, et al., 1996). Chloroquine neutralizes the acidity of lysosomes and 

inhibits lysosomal acid hydrolases, but proteolysis of HAHAB continued in the presence 

of this drug. Consistent with this, aldolase A and B are thought to undergo limited 

proteolysis by a neutral lysosomal peptidase (Horecker, et al. 1985). Though the data 

are not conclusive, past investigators have shown that this peptidase might occur on the 

cytosolic surface of lysosomes (Pontremoli, et al., 1982) which in itself could confer 

resistance of the peptidase activity to 3MA and CHQ. Given these possibilities, 3MA or 

CHQ resistant "cytosolic" proteolysis for aldolase B (HAHAB) in Figure 4-12 might 

actually occur by lysosomal mechanisms. The highest E a found published for cytosolic 

proteolysis was E a = 27 ± 5 kcal/mole reported for an enzyme now established as the 

20S proteasome (Hough et al., 1986; Bond and Butler, 1987; Waxman, et al., 1987; 

Coux, et al., 1996). E a for HAHAB proteolysis was significantly greater than this (p < 

0.019), suggesting a more complex mechanism like sequestration into lysosomes. The 

mechanism of 3MA and chloroquine resistant proteolysis of aldolase B (HAHAB) 

remains unidentified, but the results support that heat stress increases both lysosomal and 

cytosolic mechanisms for the degradation of aldolase B by direct thermodynamic 

stimulation of elevated temperatures. 

A Model For the Degradation of Aldolase B 

Figure 4-13 summarizes the results of the analysis above in a hypothetical model 

for the degradation of aldolase B. It previously was established (Chapter 1) that 

starvation for amino acids and serum (S) can induce autophagic degradation of aldolase 

A and B (Fig. 4-13, B1-»C->D). In Chapter 3, evidence was provided that starvation 

and heat stress (HS) induced ubiquitin-mediated autophagy (Fig. 4-13, 

A->B2-»C1-»C2-»D). Since it appeared that ubiquitinated aldolase B was enriched in 

isolated autophagic vacuoles and lysosomes relative to cytosol, sequestration of aldolase 

B appears to be enhanced by ubiquitination (B2>B1). Ubiquitination is known to 

mediate cytosolic proteolysis of proteins into 6-9 amino acids long peptides (Coux, et al., 
1996; Lowe, et al., 1995) indicated by A->E (lower arrow?)->F (Fig. 4-13). Though 
aldolase B occurs in ubiquitinated forms, when ubiquitination was inhibited (heat 
stressed ts20 cells) an enhance cytosolic (3MA and CHQ resistant) proteolysis of 
aldolase B still occurred (Fig. 4-13, A-»E (upper arrow)-»F). 

The proteolysis at step F increased with temperature consistent with direct 
thermal enhancement of chemical reactivity. The estimated activation energy (E a = 35- 
40 kcal/mole) was closer to a known value for a lysosomal pathway (E a = 35 kcal/mole) 
than for the major cytosolic protease, the 20S proteasome (E a = 27 kcal/mole). 
However, there was no difference from the 20S proteasome at a 0.01 significance level. 
Furthermore, a basal level of ubiquitination can be measured in ts20 cells (Gropper, et 
al., 1991), and ubiquitination is thought to be rate limiting (E a > 27 kcal/mole) for 
proteosomal degradation (Coux, et al., 1996), indicating that step F proteolysis might 
still occur in the proteasome. Previously proposed mechanisms for limited proteolysis of 
aldolase B on the cytosolic surface of lysosomes (Pontremoli, et al., 1982) and for non- 
autophagic sequestration into lysosomes (Dice and Chiang, 1988) argue that the high E a 
of step F proteolysis might still involve lysosomes. Current data is insufficient to 
definitively distinguish these possibilities. Here, cytosolic proteolysis was preferred as 
the most likely possibility. 

At higher temperatures, cytosolic proteolysis of aldolase B (Fig. 4-13, step F) 
becomes equal to or greater than sequestration into lysosomes (step B). This 




B. Sequestration 

0.3%/h |30.5°C(CT) 

{ 0.9%/h } l{ 37 °C (TCA, S) } 
1.7%/h |39.5°C(HS) 
2.9%/h | 37t(S) 

F. Proteolysis 



37°C (ForS) 
39.5° C (HS) 



G. Degradation 

1.5%/h I (all) 

amino acids 




C. Proteolysis 



D. Degradation w 





amino acids 

Figure 4-13: Pathways for Degradation of Aldolase B . Large stippled box, lysosomal 
compartments (autophagic vacuoles and lysosomes); small box, aldolase B polypeptide 
maintaining predominately native (solid outline) or denatured (broken outline) 
conformation; AldB*, full length active enzyme; AldB, proteolytically processed inactive 
enzyme; Ub-AldB, ubiquitinated aldolase B; peptides, hypothetical fragments of aldolase 
B; arrows and upper case bold face letters, processes discussed in text; %/h, median 
fractional rate of flux (SEM < 1/6 of magnitude shown, n = 24-43) through the 
indicated process at temperature and condition: CT, control temperature; HS, heat 
stress; F, fed; S, starved; "Proteolysis" indicates cleavage in to polypeptides; 
"Degradation" indicates complete proteolysis to amino acids (TCA soluble). {37°C 
(TCA, S)}, rate of autophagic sequestration at 37°C for total TCA precipitable 
polypeptides during starvation. 

interpretation was made because during heat stress aldolase B degradation was 

independent of autophagic degradation in lysosomes. Since TCA precipitable 

polypeptides were still reliant on autophagic degradation, step F was considered to 

produce peptides, consistent with involvement of the proteasome. Complete cytosolic 

proteolysis to amino acids occurred (Fig. 4-13, step G) which was fairly constant with a 

low temperature dependency, E a = 5.1 kcal /mole, that was temperature-independent at a 

0.01 significance level (Fig. 4-12, Cyto TCA). The balance of peptides produced in step 

F are sequestered (step B3) and degraded (step D2) in lysosomes. According to TCA- 

precipitable polypeptides, the autophagic degradation rate during starvation (TCA S) 

was 0.9%/h consistent with the temperature-dependency shown in Figure 4-12 (Lyso 

TCA). However, in the same cells (Fao) under the same treatment (starvation), 

lysosomal degradation of aldolase B was threefold higher, 2.9%/h (Fig. 4-13, step B, 

(S)). This indicates that lysosomal mechanisms have a preference for degrading aldolase 

B over TCA precipitated total proteins, suggesting an additional mechanism. In Chapter 

5, a role for receptor-mediated targeting to lysosomes is demonstrated for aldolase B. 



Chapter 1 introduced the hypothesis that stress-induced degradation of aldolase 
B requires ubiquitination and receptor-mediated targeting (Fig. 1-4). In Chapter 3, 
evidence was presented that aldolase B was ubiquitinated and enriched in lysosomes for 
a ubiquitin-mediated autophagic degradation during starvation and heat stress. In 
Chapter 4, temperature increase caused an increase in cytosolic and lysosomal 
degradation consistent with direct thermal stimulation of chemical reactivities. During 
starvation for amino acids and serum, autophagic degradation of total TCA precipitated 
polypeptides was consistent with the temperature-dependent rate predicted by thermal 
stimulation at 37°C (Fig. 4-13, step B). 

Lysosomal degradation of aldolase B in Fao cells was threefold higher than for 
TCA precipitated polypeptides (Fig. 5-1), suggesting a mechanism preferential for 
aldolase B sequestration into lysosomes. In Figure 5-1, the total height of each bar 
indicates the total degradation rate during starvation for the indicated proteins, including 
degradation rates for total long-lived proteins (first bar, Total), permanently expressed 
recombinant aldolase B (third bar, HAHAB), and rate of loss for endogenous aldolase B 
estimated on western blots (second bar, aldolase B). Each bar is divided into two parts. 




- 5% 


a: 4% 
1 3% 


D) 2% 





M Lysosomal 
D Cytosolic 


preci pita ted 




Aldolase B 

Protein Assayed 




Figure 5-1 : Starvation-Induced Lysosomal Degradation is faster for Aldolase B than for 
Other Long-lived Proteins . Starvation-induced degradation was measured at 37°C as 
previously described (Figs. 4-6, 4-9, and 4-10). Degradation of protein which was 
resistant to lysosomal inhibition is superimposed on the total starvation-induced rate to 
show the relative contribution of degradation (mean ± SEM, n = 8-24) occurring in 
cytosol (cytosolic) and lysosomes (lysosomal). All cytosolic degradation rates were 
comparable to each other; lysosomal degradation for Aldolase B or HAHAB was greater 
than TCA-precipitated (Student's t-test, p < 0.001). 

The lower part indicates the degradation rate contributed by proteolysis in the cytosol 
(Fig. 5-1, cytosolic, light-shaded part). These rates were calculated by inhibiting 
lysosomes during starvation. They agreed with each other and were not statistically 
different from basal degradation under fed conditions (Fig. 4-9). The results suggest that 
cytosolic degradation of aldolase B probably follows a common basal mechanism used 

by most long-lived proteins, consistent with the major role proposed for the cytosolic 

proteasome (Coux, et al., 1996). Further, this indicated that in the presence of amino 

acids (fed conditions) basal lysosomal degradation was too low to detect in Fao 

hepatoma cells, consistent with the ability of liver-derived cells to respond to amino acid 

concentrations by down-regulating protein degradation (Hendil, et al., 1990). 

In Figure 5-1, the upper part of each bar represents starvation-induced 
degradation that occurred in lysosomes (lysosomal, dark-shaded part) calculated by the 
degradative rate lost by treatment with lysosomal inhibitors. Lysosomal degradation 
accounted for the increase in protein degradation caused by starvation. During 
starvation, lysosomal degradation of total long-lived proteins (TCA-precipitated) was 
about 0.9%/hour, but lysosomal degradation of immunodetected aldolase B was much 
higher at 2.9%/h. This indicated that aldolase B utilized an additional lysosomal 
mechanism (relative to most long-lived proteins) that mediated ~2%/h greater rate of 

Intralysosomal degradation is rapid and limited by the rates of delivery of 
proteins into lysosomes. Starvation can enhance two mechanisms for delivery of 
cytosolic proteins into lysosomes: bulk uptake by autophagy and selective uptake by a 
receptor-mediated mechanism. Above and in previous chapters, autophagic mechanisms 
were considered. In this chapter, evidence demonstrates the existence of a lysosomal 
targeting signal in aldolase B that specifically mediates starvation-induced degradation. 
Lysosomal targeting signals have been characterized for other cytosolic proteins that 

follow receptor-mediated delivery to lysosomes. A description of the proposed 

mechanism for receptor-mediated targeting is presented in Chapter 1 . Aldolase B had 

characteristics of substrate proteins for this degradative mechanism, including three 

potential targeting signal motifs (Fig. 1-2). These targeting motifs were mutated, and 

starvation induced degradation of mutant protein was compared to wildtype. Below, 

glutamine #111 is shown to be essential for the starvation-induced degradation of 

aldolase B. 

Transient Expression of RABM Mutations in Putative Lysosome Targeting Signals 

In order to examine a role for three potential lysosomal targeting signals in the 
aldolase B sequence (Fig. 1-2), plasmid vectors for expressing RABM (rat aldolase B 
with a carboxyl terminal myc epitope tag) were altered by site-directed mutagenesis to 
change an "essential" glutamine (Q) in each signal to either threonine (T) or asparagine 
(N) (Table 5-1). The three "essential" glutamines occurred at defined amino acid 
positions in the aldolase B sequence: residue numbers 12, 58, and 1 1 1 (start methionine 
= 1). A three letter "mutation code" was adopted here to label the data such that the 
amino acid at positions 12, 58, and 1 1 1 were represented in order by the single letter 
amino acid convention. In this manner, QQQ indicated wildtype, and QQN indicated 
that QUI was changed to asparagine (N). 

At least seven site-directed mutant RABM's representing all mutant permutations 
needed to be expressed in cells at sufficient levels for biochemical detection. Given the 
same efficiency as for the isolation of the FaoAB cell line (one permanent expressing line 
per 340 G418-resistant Fao clones), 2,000-3,000 transfected lines would need isolated, 


Table 5-1: Site-Directed Mutations in Aldolase B Sequence for Potential 
Signal Motifs of Receptor-Mediated Targeting to Lysosomes 

Labels Used 

Amino Acid Residue** at the Essential Glutamine 

for Each of Three Potentia Motifs 





Site 2 


Site 3 







RAB or Fao 

























































*RAB, rat aldolase B; HAB, human aldolase B (97% identical to RAB; Fig. 1-2); 
HAB withl2CA5 epitope at amino terminus; RABM, RAB with 9E10 epitope at C- 
** Single letter convention 

amplified, and screened for permanent expression. To avoid this, other candidate cell 
lines were screened for transfection efficiency (Table 5-2). Transfection efficiency was 
defined as the fraction of cells expressing RABM by immunofluorescence screening. 
Some commercial transfection lipid formulations were also screened and found that 

transfection lipids operated differentially between cell lines. For example, Pfx-5 

appeared to effectively transfect BHK cells but in other cell lines was toxic, produced an 

inclusion body artifact, or failed to express RABM (Table 5-2). 

The HuH7 human hepatoma cell line transfected with Pfx-3 (Invitrogen) had the 

highest relative transfection efficiency for expressing RABM (Table 5-2), so isolation of 

HuH7 cell lines permanently expressing RABM was attempted. Hundreds of HuH7 

colonies isolated after transient transfection acquired resistance to the selectable marker 

(G418), but none of forty clonal cell lines retained sufficient expression of recombinant 

aldolase B for biochemical studies. This was consistent with the low incidence of 

permanent expression of recombinant aldolase B in Fao cells, indicating a need for 

massive screenings. However, the transient 

Table 5-2: Optimization of Cell Line and Lipid Type 
for Transient Expression of RABM* 

Cell Lines 



















































"%, fraction of cells immunofluoresently labeled for RABM-specific expression; 
toxic, >50% cells had wrinkled appearance, obvious sloughing off present; 
inclusions, large fluorescent inclusion bodies with little or no cytosolic labeling. 

expression of RABM in HuH7 cells was relatively high and after optimization appeared 
sufficient for biochemical assays (Table 5-3). Commonly, protocols call for plating cells 
the day before transfection, but isolated individual HuH7 cells in cultures one day after 
plating were killed by Pfx-3 treatment (data not shown) which limited transfection 
efficiency (Table 5-3). By incubating cells for two days after plating, most cells 
contacted other cells, survived Pfx-3 mediated transfection, and transfection efficiency 
was increased by tenfold (Table 5-3). Though not generally examined during 
optimization of transfection, the incubation time between plating cells and transfection 
was found to have a large effect on transfection efficiency. Consistent transient 
transfection efficiency of >15% (Table 5-3) indicated that transient transfection could 
replace the more laborious isolation of permanent cell lines. 

Table 5-3: Effects of Timing and LipidrDNA Ratio on 
Transfection Efficiency for RABM Expression in HuH7 Cell Line 

Time After Plating 

Transfection Time 

Pfx-3 LipidrDNA Ratio 
(DNA held constant) 





















In order to compare degradation rates for wildtype and mutant RABM, 
expression levels for the mutants had to be sufficient. HuH7 cell transfection efficiencies 

Figure 5-2: Transient Expression of Recombinant Aldolase B in HuH7 Cells . HuH7 
human hepatoma cells were transfected for optimum transient expression of various 
epitope-tagged aldolase B proteins defined in Table 4-1, processed for epitope-tag 
specific immunofluorescence microscopy (9E10 for all, except 12CA5 for panel b), and 
micrographs exposed until background labeling just became apparent; panels a through e, 
representative fields showing labeling patterns for recombinant aldolase B proteins 
defined in Table 5-1, similar fields occurred with expression of all the different proteins 
except RABM's with threonine (T) at position #58; panel f, expression of QTQ mutant 
RABM showing the brightest cell on an entire coverslip, only rare cells were clearly 
brighter than background but dim out-lining of nuclei was visible in many cells that did 
not occur on untransfected control coverslips (not shown). Scale bar = 50 u,M. 


were consistently greater than 15% for RABM proteins (Fig. 5-2, panels a-f) using 

wildtype (QQQ) and site-directed mutants with changes in amino acid residues #12 or 

#111 (TQQ, QQT, QQN, and TQT). Such high levels of expression were sufficient for 

isolation by immunoprecipitation which confirmed comparable expression levels between 

wildtype RABM and these mutants (Fig. 5-3). Note, the transiently expressed proteins 

were similar for different mutants and wildtype RABM (41.3 kD) and did not occur in 

cells transiently expressing (3-galactosidase ((3Gal). Previously during permanent 

expression, RABM became associated with endogenous 40 kD aldolase which could be 

co-immunoprecipitated (Fig. 4-8b). However, immunoprecipitation from transiently 

transfected cultures pelleted very little or no 40 kD protein, indicating that transient 

protein expression per cell was so high relative to endogenous aldolase expression that 

the exogenous RABM predominately associated with itself presumably forming homo- 

oligomeric complexes of RABM. Since transient overexpression excluded association 

with endogenous aldolase, possible complications from oligomeric interactions with 

endogenous aldolase subunits were reduced. For example, if all subunits of an aldolase 

tetramer are degraded together, then wild type signals in an endogenous subunit could 

rescue an associated mutant subunit. Transient expression seems to eliminate this 

possibility. The results indicated that wildtype RABM and mutants at residues #12 and 

#111 were expressed sufficiently for protein degradation measurements. 

Replacing glutamine #58 with threonine (QTQ, TTQ, QTT, or TTT) prevented 

high RABM expression (Fig. 5-2, panel f), and further biochemical analysis could not be 









Aldolase B 






a a a 







Figure 5-3: Immunoprecipitation of RABM and its Mutants Transiently Expressed in 
HuH7 Cells . HuH7 cell cultures were transfected for transient expression of 0- 
galactosidase (PGal) or RABM forms identified by the mutation codes shown in Table 5- 
1; cells were then metabolically labeled with 35 S-methionine then processed for anti-myc 
epitope (left panel, 9E10) or for anti native aldolase B (right panel, Aldolase B) 
immunoprecipitation as described in Materials and Methods; autoradiographs of dried 
gels of immunoprecipitation pellets run on SDS-PAGE are shown; left and right panels 
are lined up on a 44.5 kD heavy background protein (actin?), and differences in 
electrophoretic migration between gels are indicated by relative positions of the 41.3 kD 
band relative to the 44.5 kD band. 

done (summarized later in Table 5-4). The minimal change in the cDNA (2 base pairs) is 

unlikely to affect transfection efficiency. Consistent with this, many very low expressing 

cells were apparent by very dim but visible negative nuclei suggesting that transfection 

occurred, but the protein was not expressed well. To get around this low expression, a 

less harsh change of glutamine #58 to asparagine (N) was tried which successfully 

produced enough protein for degradation assays (Figs. 5-2, panel d, QNQ). In the QNQ 

mutant, an amide side chain chemistry was retained at residue #58, suggesting a possible 

role in aldolase B protein stability. 

Other evidence supports a role for glutamine #58 in the structural stability of 
aldolase B. First, glutamine #58 is embedded in the tertiary structure of the protein, 
placing it in the right place for such a role. Glutamine #58 is not a highly conserved 
residue which argues against this role. However, neighboring amino acid sequence was 
examined in different mammalian aldolase isoforms and found that when glutamine #58 
was missing then glutamine occurred at a position three residues away. These glutamine 
side chains would occur on the same side of an a-helix in a similar 3-D position and 
could perform the same structural role. Supporting this idea, glutamine #58 occurs in an 
a-helical region of aldolase structure. 

The results here provided wildtype and mutant RABM proteins transiently 
expressed in HuH7 cells at similar levels (Figs. 5-2 and 5-3), allowing effects on 
starvation induced degradation to be examined for mutations at all three potential signal 
sequence motifs found in aldolase B. 

Starvation Induces Autophagic Degradation in HuH7 Cells 

Before examining transiently expressed aldolase B (RABM), starvation-induced 
degradation of total long-lived proteins in HuH7 cells was examined. Chloroquine 
(CHQ) but not 3-methyladenine (3 MA) significantly reduced basal degradation of long- 
lived proteins in HuH7 cells (Fig. 5-4). This suggested that a fraction of unstressed basal 
protein degradation occurred in lysosomes but not by 3MA-sensitive autophagy. Both 

« 2.0% 

* 1.5% ■■ 

~ 1.0% 1 




20 uM CHQ 40 uM CHQ 

Inhibitor Treatment 



Figure 5-4: Starvation-Induced Autophagic Degradation of Long-lived Proteins Occurs 
in HuH7 Cells . HuH7 cultures were metabolically labeled with 14 C-valine, chased with 
10 mM "cold" valine in DMEM+10% FBS (Fed) or Krebs-Heinseleit without amino 
acids (starved), and loss of TCA precipitable counts was used to calculate degradation 
rates (mean ± SEM, n = 12) as described in Materials and Methods; media were 
unsupplemented (control) or supplemented with indicated concentrations of chloroquine 
(CHQ) or 3-methylyadenine (3MA); Student's t-tests: inhibitor treated < corresponding 
control degradation rate (*, p < 0.013); Starved > Fed degradation rate (**, p < 0.01; 
***, p < 0.005, ****, p < 0.0001). 

(Starved) and basal (Fed) degradation of long-lived proteins was inhibited with 
chloroquine (CHQ) such that starvation appeared to induce a mechanism that was 
resistant to CHQ (Fig. 5-4). Alternatively, the results were consistent with mechanisms 
for fed and starved degradation having a similar sensitivity to CHQ. Generally, 
lysosomal degradation plays a greater role in starvation-induced degradation (Starved) 
than in basal degradation (fed). Since the results were inconsistent with this, HuH7 cells 
might have a non-specific sensitivity to CHQ which affects both basal and starvation- 
induced degradation equally. Doubling inhibitor concentration did not change results 
(Fig. 5-4), indicating that a maximal effect occurred, but the possibility that lower CHQ 
concentrations might be more specific for starvation-induced degradation in HuH7 cells 
was not examined. 3-methyladenine specifically reduced the enhanced degradation of 
total long-lived proteins by about 0.8%/h in HuH7 cells (Fig. 5-2), consistent with the 
0.9%/h inhibition of enhanced autophagic degradation in Fao cells (Fig. 4-6). The 
results suggested that starvation-induced autophagic degradation of total long-lived 
proteins occurred in HuH7 cells. 
Transient Expression Does Not Affect Starvation-Induced Degradation of RABM 

I showed that HuH7 cells undergo starvation-induced degradation of long-lived 
proteins and that 9E10 myc-epitope tagged rat aldolase B (RABM) could be transiently 
expressed at sufficient levels for biochemical analysis. In this section, the degradation of 
the transiently expressed RABM is shown to be comparable to permanently expressed 
recombinant aldolase B (HAHAB) and endogenous aldolase B of Fao cells. 

I found that variation in permanent expression levels had little or no effect on 

degradation rates in E36 cells (data not shown), but a sub-population of the transiently 

transfected HuH7 cells clearly had higher RABM levels per cell than previously 

examined. Thus, it was important to establish whether transient transfection altered the 

degradation of 





6% - 


D Fed m Starved 


(endogenous) (permanent) (transient) (transient) 

Fao FaoAB HuH7 HuH7 

Aldolase Form (expression) & Cell Line 

Figure 5-5: Differences in Expression and Epitope Tag Have No Effect on Starvation- 
Induced Degradation of Aldolase B in Hepatoma Cell Lines : Cultures of cell lines 
indicated expressing aldolase B were measured for fed and starved aldolase B 
degradation rates ± standard error (n = 25 to 37) according to protocols compatible for 
each form of the enzyme as described in Materials and Methods. RAB (endogenous) 
Fao, endogenous rat aldolase B in Fao rat hepatoma cells; HAHAB (permanent) Fao, 
permanently expressed amino HA-tagged human aldolase B transfected in Fao cells; 
RABM (transient) HuH7, transiently expressed carboxyl Myc-tagged rat aldolase B in 
HuH7 cells; HAHAB (transient) HuH7, transiently expressed HAHAB in HuH7 human 
hepatoma cells. Student's t-test: Starved > Fed (*, p < 0.0005); HAHAB (Starved) in 
HuH7 > in FaoAB and RAB (starved) in Fao > RABM in HuH7 (**, p < 0.001) 


aldolase B. Figure 5-5 compares the basal and starvation-induced degradation of various 
forms of aldolase B expressed in various ways. The results are similar regardless of the 
variations. There was little or no effect caused by epitope tagging. Tagging at the 
amino terminus or the carboxyl terminus gave identical results, and using 12CA5 versus 
9E10 epitope also had no effect. Whether endogenously expressed, permanently 
expressed from transfected cDNA, or transiently expressed as described above the 
results were similar (Fig. 5-5). 

Site-Directed Mutations Did Not Affect Wildtype Activity of RABM 

Endogenous 40 kD aldolase was absent in proteins immunoprecipitated from 
transiently transfected cells (Fig 5-3), indicating that characteristics of 
immunoprecipitated RABM could be examined in isolation from endogenous aldolase 
effects. Antiserum against aldolase B also failed to pull down endogenous aldolase 
indicating that the endogenous isoform of HuH7 cells is probably aldolase A (Fig. 5-3). 
These results indicated that exogenous RABM is for the most part separate from 
endogenous aldolase, during transient expression. 

Even though endogenous aldolase was not associated with transiently 
overexpressed RABM, a 42 kD protein was specifically immunoprecipitated from 
RABM expressing but not P-galactosidase expressing cells (Fig. 5-3). This indicated 
that RABM associated with an endogenous 42 kD protein. Aldolase isozymes have an 
established binding affinity for 


glyceraldehyde-3 -phosphate dehydrogenase (GAPDH) which has a molecular weight of 
42 kD. The results suggest that RABM retains GAPDH binding activity. If so, then all 
the mutants examined had no effect on GAPDH binding, because the 42 kD protein was 
co-immunoprecipitated with them as well (Fig 5-3). 

With transient expression, immunoprecipitation was used to isolate epitope- 
tagged aldolase B away from endogenous aldolase, giving purified RABM or HAHAB. 
By measuring activity specifically precipitated by the epitope tag, the effect of site- 
directed mutations on the enzymatic activity of aldolase B could determine. As shown in 
Figure 5-6, all forms of recombinant aldolase B tested had similar activity regardless of 
epitope tag or site-directed mutations. Thus, effects of mutations on starvation-induced 
degradation shown below are not due to indirect effects of altered aldolase B activity 
(Fig. 5-6). 

Glutamine Residue #111 is Required for Starvation-Induced Degradation of Aldolase B 

Above, the degradation of RABM transiently expressed in HuH7 cells was shown 
comparable to the degradation of permanently expressed HAHAB and endogenous RAB 
(Fig. 5-5). These aldolase B forms followed a starvation inducible pathway for 
degradation. Then mutant forms of RABM were transiently expressed in HuH7 cells; 
the different mutations are described and defined in Table 5-1. 




12 3 4 5 6 7 

Protein Expressed 

HAHAB beta-Gal 

QQQ (12CA5) 

8 9 

Figure 5-6: Different Forms of Recombinant Aldolase B Retain Similar Levels of 
Catalytic Activity . Replicate HuH7 cell cultures were transfected for transient 
expression of indicated proteins, lysed in 1% NP40/1 mM EDTA/TBS pH 7.5, and 
immunoprecipitated for 9E10 (bars 1 through 7) or 12CA5 (bars 8 and 9) epitopes as 
detailed in Materials and Methods under subsaturating conditions. After standard 
washes, pellets were washed with TE pH 7.5 then suspended in 1 ml aldolase reaction 
mixture. Periodically, each sample was pelleted and OD(340 nm) of supernatant was 
measured. The mean ± SEM (n = 3) OD decrease caused by adding FDP was 
normalized to the non-specific background of cultures expressing P-galactosidase and 
immunoprecipitated with the same antibody (bars 7 or 9). Student's t-test: relative 
activity for RABM or HAHAB expressing cultures > untransfected non-specific 
controls, lanes 7 and 9 (p < 0.0001). 

Figure 5-7 shows that glutamine #111 is essential for starvation-induced 
degradation of RABM. Changing at glutamine #12 to threonine (TQQ) yielded fed and 
starved degradation rates for RABM that were statistically indistinguishable from 
wildtype aldolase B sequence (QQQ). Even though the peptide sequence at glutamine 
#12 is the most conserved of the mutated signals (identical in all vertebrate aldolases), it 

did not function as a signal for starvation-induced degradation (Fig. 5-7). A role for 

glutamine #58 could not be assessed by threonine substitution, because this resulted in 

expression too low for degradation assays. When glutamine was replaced with 

asparagine at residue #58 (QNQ), conserving an amide at this residue, expression was 

close to other high level transient expressions, and starvation-induced degradation was 

like wildtype (Fig. 5-7). The results indicated that glutamines at residues #12 and #58 

were not essential for starvation-induced degradation and suggested that the lysosomal 

targeting motifs at these positions were not functional in vivo. 

Whenever the glutamine at residue #111 was mutated then starvation-induced 

degradation of RABM was completely inhibited (Fig. 5-7). For wildtype aldolase B 

sequence in RABM (QQQ), starvation increased degradation of RABMQQQ by greater 

than twofold, but when glutamine residue #111 was changed to asparagine (QQN) or 

threonine (QQT and TQT), the resulting mutant RABM failed to show an increase in its 

degradation in response to starvation. Since the loss of inducible degradation occurred 

even with a minimum change of glutamine to asparagine (QQN), glutamine #111 is 

clearly essential for starvation-induced degradation of aldolase B and asparagine cannot 

function in its place. This is consistent with the function of a motif originally described 

by J. Fred Dice and indicates that aldolase B is a substrate for receptor-mediated 

targeting to lysosomes. 



RABM Form 


Figure 5-7: Glutamine Residue #1 1 1 is Required for Starvation-Induced Degradation of 
Aldolase B : Various forms of RABM were transiently expressed in HuH7 cells, then fed 
and starved degradation rates ± standard error (n = 18 to 33) for RABM were measured 
as described in Materials and Methods. Different forms of RABM are labeled according 
to the mutation code given in Table 5-1. Student's t-test: *, Starved > Fed, p<0.0001; 
**, mutant < wildtype = QQQ, p < 0.0001. 

Glutamine #111 Specifically Mediates Starvation-Induced Degradation 

Above, it was demonstrated that changing glutamine #111 of aldolase B affects 
its starvation-induced degradation. Here, it is proposed that this change directly affects a 
degradative mechanism, but if some other function of this enzyme is blocked, then 
inhibition of starvation-induced degradation might be an indirect affect of other lost 
function. To address this, evidence that various aldolase functions were similar between 

the RABM containing wildtype sequence (QQQ) and mutants at glutamine #111 (QQN, 

QQT, and TQT) was collected. 

Table 5-4 summarizes evidence that mutations at glutamine #111 were specific 

for blocking the starvation-induced degradation of aldolase B. First, expression levels 

were similar to wild type RABM, including transfection efficiencies (Table 5-4, column 

2) and immunoprecipitated RABM (Fig. 5-3). Second, aldolase is known to bind f-actin 

(O'Reilly and Clarke, 1993; Wang, et al., 1996), and HuH7 cells have well developed 

stress fibers that became decorated with antibodies specific for RABM (Fig. 5-8, 9E10). 

Localization of RABM to actin was confirmed by double labeling with phalloidin which 

specifically labels f-actin (Figure 5-8b). For all mutant RABM's used in degradation 

experiments, similar stress fiber morphology (Stress Fiber Label) was observed as with 

RABM-specific label in RABMQQQ wildtype expressing cells (data not shown). This 

provided evidence that some actin-binding activity was retained in these mutants (Table 

5-4, column 3) Third, aldolase is known to bind GAPDH (Clarke et al., 1982), and 

when mild conditions were used for immunoprecipitations (Fig. 5-3, Aldolase B) a 42 

kD protein was specifically pulled down. Though its identity was not confirmed, the 

sized matched that predicted for the 42 kD subunits of GAPDH (Clarke et al., 1982). 

The results suggest that GAPDH binding was retained by site-directed mutants of 

RABM (Table 5-4, column 4). Above, enzymatic activities were already shown to be 

comparable to each other regardless of the mutations (Fig. 5-6, Table 5-4, column 5). 

Lastly, there were no significant differences between various basal degradation rates for 

all the assayed RABM forms, including the mutants. Furthermore, mutation of 


Table 5-4: Summary of RABM Forms Expressed in HuH7 Cells 


Form 1 



(% Brightly 

Labeled Cells) 

Label 2 

Binds to 
protein 3 

Activity 4 

of RABM 5 

of RABM 6 











































Mutants Expressed too Low for Additional Characterization 1 

^n/d, not done) 




































1 See mutat 

on codes in Table <■ 


+ : common, often bright; ± : occasional, usually dim, observation based on rare labeling 


specific co-precipitation of an endogenous 42 kD protein (probably GAPDH), see Fig. 4-28 

+ : activity in immunoprecipitate lacking 40 kD protein (endogenous aldolase), see Fig. 4-29 

5 + : basal degradation = published values (1.0-2.5%/h); see Fig. 4-31 

6 + : starved degradation > basal; - : starved degradation = basal; see Fig. 4-3 1 

glutamine #111 specifically inhibited the induced portion of RABM degradation, 
indicating that the basal degradation of mutated RABM's continued during starvation. 
The data support that the basal degradation of RABM continued without being affected 
by the site-directed mutations (Table 5-4, column 6). Together the results demonstrate 
that glutamine #111 in aldolase B is likely to play a direct role in receptor-mediated 
targeting to lysosomes induced by starvation. 




Figure 5-12: RABM has Actin Binding Activity , a) left panel, BHK (baby hamster 
kidney) and right panel, HuH7 cells were transfected for transient RABM expression and 
single labeled for 9E10 immunofluorescence microscopy as described in Materials and 
Methods, field showing example of occasional cells having good stress-fiber labeling 
indicated by arrows; b) HuH7 cells were labeled as in (a) and then double stained with 
phalloidin carrying an alternative fluorophore, the same field is shown: left panel, the 
9E10 fluorescence channel; right panel, the phalloidin fluorescence channel. 

At the end of Chapter 3 and the beginning of this chapter, starvation-induced 

degradation of aldolase B was examined in hepatoma cell lines. It was confirmed that 

starvation-induced degradation occurs by autophagic degradation for long-lived proteins 

in general including aldolase B. I found a residue, glutamine #111 that was specifically 

essential for starvation-induced degradation of aldolase B, indicating that it served as 

part of a degradative targeting signal. Together, the results of Chapters 3 through 5 

support the hypothesis that stress-induced degradation of aldolase B requires 

ubiquitination and a receptor-mediated targeting signal. 



A model was presented showing two mechanisms known to mediate stress- 
induced degradation of long-lived proteins (Fig. 1-4). In Figure 6-1, the findings of this 
study were used to update the initial model with aldolase B utilized as a representative 
long-lived cytosolic protein. Evidence indicated that autophagy, ubiquitination, cytosolic 
proteolysis, and a molecular recognition signal function in the degradation of aldolase B 
in response to heat stress and starvation. 

Autophagy and Ubiquitination 

This study began by identifying ubiquitinated forms of aldolase B with antibodies 
raised against ubiquitin-free aldolase B. In Figure 6-1, ubiquitination is represented by 
process 6 which maintains aldolase B partially denatured as indicated by preferential 
reactivity with antibodies to denatured epitopes (Fig. 3-2). Unmodified aldolase B 
predominated in cytosol, whereas ubiquitinated aldolase B was enriched in lysosomal 
compartments, especially autophagic vacuoles (Fig. 1-3 and 3-3). Furthermore, 
lysosomal association of a major ubiquitinated form of aldolase B (Ub68) increased 
during amino acid starvation and decreased when autophagy was inhibited. Ub68 also 
occurred in E36 cells that undergo increased autophagic degradation during heat stress 


Figure 6-1: Mechanisms for Stress-Induced Degradation of Cytosolic Proteins in 
Lysosomes . Autophagy (upper pathway) and receptor-mediated targeting (lower 
pathway) were proposed for stress-induced delivery of cytosolic proteins to lysosomes 
for degradation; the arbitrary cytosolic protein is shown as a tetramer (aldolase B occurs 
as a tetramer); components of the pathways are labeled on the diagram; processes are 
labeled by boxed numbers: la & b, association with or engulfment by pre-autophagic 
membranes; 2, sequestration into double-membrane bound autophagic vacuole; 3, 
maturation of autophagic vacuole (acidification and acquisition of lysosomal hydrolases); 
4a & b, proteolysis into polypeptide fragments; 5, complete degradation to amino acids; 
6, disassembly and denaturation of structure by an unknown factor; 7, association with a 
receptor complex on the lysosomal surface; 8, translocation across the lysosomal 
membrane; 9, limited proteolysis and removal of ubiquitin; 10, proteolysis into 
polypeptide fragments; 1 1, complete degradation to amino acids. 

(Fig. 3-11). In response to heat stress, endogenous aldolase A of E36 cells accumulated 
in lysosomes dependent upon elevated levels of ubiquitination (Fig. 3-5). These results 

suggested that ubiquitination facilitated the autophagic delivery of aldolase B to 

lysosomes (Fig. 6-1, process lb > la), but other data indicate that autophagic 

degradation is affected at a later step in the pathway. 

Ubiquitination was previously proposed to mediate degradation in 
autolysosomes (Fig. 6-1, process 4) rather than sequestration, because autolysosomes 
expanded sixfold in volume when ubiquitination was inhibited (Lenk et al., 1992). 
Consistent with this, lysosomal degradation was temperature-dependent (Fig. 6-12), 
increased approximately 6-fold during heat stress relative to control temperature (Fig. 6- 
13), and required ubiquitination (Figs. 3-14 and 3-15). Additionally, ubiquitination was 
required for limited proteolytic processing of an epitope-tagged form of aldolase B 
associated with lysosomes (Fig. 3-10). This evidence supports a role for ubiquitination 
in proteolysis mediated by at least some lysosomal proteases (Fig. 6-1, process 9). 

Whether for delivery or degradation, this study demonstrates that ubiquitination 
mediates lysosomal mechanisms for degradation of aldolase B. The hypothesis that 
aldolase B requires ubiquitination for enhanced degradation in lysosomes was supported. 
Given this, a number of questions arise. What must be ubiquitinated, aldolase B or a 
component of the autophagic mechanism? Is ubiquitinated aldolase B sequestered or 
degraded faster than unmodified aldolase B? What does ubiquitination do to the 
structure and function of aldolase B? 

Preliminary evidence indicates that ubiquitinated aldolase B is maintained in a 
more denatured conformation. Does ubiquitination mediate disassembly of aldolase 

tetramers? Does it expose potential targeting signals for degradative mechanisms? Is 

ubiquitinated aldolase B preferentially degraded? Ubiquitinated aldolase B constitutes at 
least part of Ub68, one the most abundant protein-ubiquitin conjugate species 
characterized on SDS-PAGE. Though enriched in lysosomes relative to total proteins, a 
significant amount of Ub68 still occurs in the cytosol (Fig. 1-3 and 3-3). Does aldolase 
B make up most or all of Ub68? How abundant is Ub68 relative to unmodified aldolase 
B? Is Ub68 a necessary intermediate for ubiquitin-mediated degradation of aldolase B? 
Could ubiquitinated aldolase B perform another function? 
Clues from Temperature-Dependent Cytosolic Proteolysis and Lysosomal Degradation 
While examining heat stress-induced autophagic degradation, a second 
mechanism for aldolase B degradation during heat stress was identified. The degradative 
activity was resistant to treatments that inhibited autophagy and lysosomal degradation, 
suggesting a cytosolic mechanism (Fig. 6-1, processes 10 and 11). Temperature- 
dependency of this cytosolic mechanism was consistent with thermal enhancement of a 
rate-limiting reaction. Calculations of the activation energy (E a ) were limited by use of 
only three temperatures (Fig. 4-12), but the large magnitude was consistent with 
regulated degradation by the complete 26S proteasome complex. An initial E a estimate 
was 28 kcal/mole (data not shown), but when more data were included E a was 35-40 
kcal/mole. This was higher than predicted for proteolysis by isolated 20S proteasomes 
(27 kcal/mole), supporting the concept that a rate limiting step for proteasomes occurs 
before 20S-mediated proteolysis by 19S regulatory components of the entire 26S 

proteasome complex (Coux, et al., 1996). However, E a 's estimated for cytosolic 

proteolysis of aldolase B could not distinguish between E a 's predicted for autophagic 

degradation (35 kcal/mole) and regulated proteolysis in 26S cytosolic proteasomes 

(presumably >27 kcal/mole). The inhibitors used here measurably blocked lysosomal 

degradation of total TCA proteins, indicating that a cytosolic protease, consistent with 

26S proteasomes, mediated proteolysis of aldolase B. 

Stress-induced degradation in proteasomes is known to be regulated by 
ubiquitination, but when ubiquitination was inhibited, cytosolic proteolysis of aldolase B 
continued. However, inhibition of ubiquitination was incomplete (Gropper, et al., 1991) 
and could not eliminate the possibility that cytosolic proteolysis continued at low levels 
of ubiquitination. Consistent with this, similar amounts of Ub68 (ubiquitinated aldolase 
B) occurred in cells having a predicted fourfold difference in ubiquitination activity (Fig. 

During heat stress, lysosomal inhibition did not affect enhanced cytosolic 
proteolysis of aldolase B (Figs. 4-4 and 4-11) but effectively blocked enhanced 
degradation of total long-lived TCA-precipitable peptides (Fig. 3-15). If aldolase B is 
representative of long-lived housekeeping proteins, then enhanced cytosolic proteolysis 
of aldolase B (loss of immunoreactivity) was limited to production of undetected TCA 
precipitable polypeptides, and enhanced degradation of total proteins (loss of TCA 
precipitability) proceeded to amino acids in lysosomes via autophagy. The results are 
consistent with a role for proteasomes in aldolase B degradation, because cytosolic 

proteasomes produce small polypeptides rather than amino acids (Lowe et al., 1995; 

Coux, et al.,1996), whereas lysosomes completely degrade proteins to amino acids 

(Hershko and Ciechanover, 1982; Mortimore and Poso, 1987; Olson et al., 1990). 

Autophagic degradation of TCA precipitated proteins and partial cytosolic 
proteolysis of aldolase B correlated with temperature such that their rates correlated with 
each other (Fig. 4-12). This suggested that both mechanisms could play a role in heat 
stress-induced degradation of aldolase B by thermodynamic stimulation of molecular 
machinery. Cytosolic partial proteolysis was faster than autophagic degradation under 
similar conditions (Fig. 4-13), suggesting the possibility that a product of the former 
might stimulate the latter. Are lysosomes regulated by products of proteolysis? Amino 
acids (lysosomal products) have been established to inhibit autophagic degradation. Is it 
possible that peptides produced by cytosolic proteases like the proteasome have an 
opposite effect, stimulating autophagy? 

Endogenous inhibitors of the proteasome have been proposed to prevent 
excessive proteolysis of cellular proteins (Coux, et al., 1996). One of the best 
characterized endogenous inhibitors, CF-2, was found to be identical to 5-aminolevulinic 
acid dehydratase (AADH), an essential enzyme in heme biogenesis (Guo, et al, 1994). 
AADH is a long-lived oligomeric housekeeping protein with 40 kD subunits that occur 
as a 50 kD ubiquitinated form associated with 26S proteasomes (Guo, et al, 1994). 
Above, aldolase B was described as a long-lived oligomeric housekeeping protein with 
40 kD subunits that occur as a 68 kD ubiquitinated form. The parallel suggests that 

aldolase B might be another endogenous proteasome inhibitor. Consistent with this, the 

proteasome has a stable intermediate during carboxyl to amino terminal proteolysis of 

substrates (Lowe et al., 1995), and aldolase B has a protease- sensitive carboxyl terminus 

(Chapter 1). The proteasome might begin proteolysis on the accessible carboxyl 

terminus but be blocked by the protease-resistant structure for the rest of aldolase B. 

Stress-induced ubiquitination of aldolase B could relax its structure, allow it to degrade, 

and thereby make the proteasome available to degrade other proteins. Though the 

identity of the cytosolic protease involved in the degradation of aldolase B still needs to 

be established, a role for the proteasome seems important to examine. 

Signal-Mediated Targeting 

Aldolase was found to have properties similar to substrates of receptor-mediated 

targeting to lysosomes (Fig. 6-1, processes 7 and 8), including three potential targeting 

signals in the aldolase B sequence (Fig. 1-2). Site-directed mutagenesis was used to 

replace glutamines defined as "essential" for each of these motifs. Of these, only 

glutamine #111 was shown to specifically mediated starvation-induced degradation of 

aldolase B. Whether aldolase B utilizes the same mechanism as previously described for 

receptor-mediated targeting to lysosomes was not determined, but the data are 

consistent with this mechanism. Previously, receptor function for this pathway was 

demonstrated in cultured cells (Cuervo, et al., 1996), but signal function had not. If 

glutamine #111 mediates this mechanism for aldolase B, then this study provides the first 

evidence for signal function in living cells. 

A major weakness of receptor-mediated targeting to lysosomes was recently 

published (Gorinsky, et al., 1996). Conformations of known substrates for this pathway 

would prevent receptor recognition, prompting investigators to propose unknown 

factors to create more extended conformations (Gorinsky, et al., 1996). Based on 

native and denatured immunoreactivities, ubiquitinated aldolase B retained a more 

denatured conformation than unmodified aldolase B. This suggested that ubiquitin can 

function as the unknown factor of Figure 1-1 to relax the structure of aldolase B and 

improve lysosomal targeting signal availability (Fig. 6-1, process 6 and 7). 

While characterizing the signal-mediated degradation of aldolase B, rat and 

human aldolase B were expressed in rat and human cells. Maximal degradation rates 

occurred when the source animal species of the aldolase B matched the species of the 

cell (Fig. 5-5, RAB in Fao and HAHAB in HuH7). When species did not match 

(HAHAB in Fao and RABM in HuH7), starvation-induced degradation rates were 35- 

40% lower (Student's t-test, p < 0.001). This indicated that species specific differences 

between rat and human aldolase B might mediate recognition by proteolytic machinery. 

Between rat and human aldolase B, there are seventeen non-identical amino acid 

residues; only twelve alter side chain chemistry (Fig. 1-2). These residues are not near 

the primary amino acid sequence for the signal at glutamine #111, indicating that some 

might function as an independent recognition signal for cellular degradative mechanisms. 

Though more work is needed to confirm these initial findings, the ability of cells to 

distinguish between rat and human isoforms would further support signal-mediated 


targeting of aldolase B during starvation. This study supports the hypothesis that a 

molecular signal mediates stress-induced mechanisms for degradation of aldolase B. 
Present and Future Contributions to the Field of Protein Turnover 

This study demonstrates that Aldolase B is a substrate for multiubiquitination and 
ubiquitin-mediated proteolysis in lysosomes. This is consistent with the hypothesis that 
ubiquitination of substrate proteins facilitates degradation in lysosomes (Lenk, et al., 
1992). A secondary hypothesis is that ubiquitination is required for a subpopulation of 
lysosomal hydrolases. In support of this, limited proteolysis of aldolase B in lysosomes 
was blocked when ubiquitination was inhibited (Fig. 3-10), and aldolase B did not 
significantly accumulate in lysosomes, indicating other proteolytic activities continued. 
In the future, the recombinant aldolase B (RABM) can be used as a marker for ubiquitin- 
dependent limited proteolysis in lysosomes. As such, limited proteolysis of RABM could 
serve as an assay for this activity, allowing purification and identification of specific 
lysosomal components that require ubiquitination. 

When aldolase A was examined, heat stress cause greater than twofold 
accumulation of aldolase enzyme activity in lysosomes, and this accumulation was 
ubiquitin-dependent (Fig. 3-5). This indicated that ubiquitination plays a role in 
sequestration of aldolase A. In support of this interpretation, ubiquitinated forms of 
aldolase B (Ub68) were enriched more in autophagic vacuoles than in cytosol or 
lysosomes (Fig. l-3a). Together, the data suggest that ubiquitination mediates at least 
two processes during lysosomal degradation: sequestration and lysosomal proteolysis. 

Autophagic sequestration in starved mammalian cells was proposed to occur by 

bulk non-selective uptake (Kopitz et al., 1990). However, in starved liver cells, 

ubiquitinated aldolase B was enriched in lysosomes, while unmodified aldolase B was 

enriched in cytosol (Fig. 3-3). Assuming ubiquitination occurs in cytosol, the results 

contradict a non-selective mechanism, and suggest that ubiquitinated proteins can 

undergo selective sequestration. Selective autophagy occurs in yeast (Tuttle et al., 

1993). Aldolase B might provide valuable evidence for ubiquitin-mediated selective 

autophagy. The different SDS-PAGE mobility of 40 kD unmodified and 68 kD 

ubiquitinated (Ub68) aldolase B can be used to separate them after co- 

immunoprecipitation (Fig. 3-11). In pulse-chase experiments, the relative loss of 

radiolabel from these two forms could be compared with and without autophagic 

inhibition. The results would determine relative degradation rates for aldolase B and 

Ub68 and their relationship with autophagic mechanisms. If ubiquitination of aldolase B 

makes it a better substrate for autophagy, then autophagic degradation of Ub68 would 

be more rapid than for 40 kD aldolase B. 

Amino acids have long been known to inhibit autophagic degradation, but how 

their levels are detected and translated into an inhibitory signal is unknown. 

Temperature-dependence for cytosolic proteolysis of aldolase B and lysosomal 

degradation of long-lived proteins were parallel with statistically equal activation 

energies (Fig. 4-12). Equal activation energies are predicted for processes controlled by 

the same rate limiting mechanism. If cytosolic proteolysis of long-lived proteins is partial 

as proposed for aldolase B above, then cytosolic peptides would be produced. If these 

peptides induce autophagy, then autophagic degradation would be dependent on the 

rate-limiting step for cytosolic proteolysis and give the same E a . This suggests a model 

in which amino acid deprivation induces limited proteolysis which produces peptides 

which induce autophagy. Autophagic degradation in lysosomes produces amino acids 

that might inhibit cytosolic proteases that produce peptides, causing peptide levels and 

peptide-dependent autophagy to be reduced. Such a model could be tested by 

introducing peptides into cytosol with liposomes, electroporation, microinjection or 

overexpression from plasmid vectors then measuring autophagic degradation. If the 

model is correct, then peptides should induce autophagy even in the presence of 

inhibitory amino acids. 

As discussed previously, the proteasome is the major protease in cytosol, is 

known to produce peptides, and might mediate cytosolic proteolysis of aldolase B. To 

examine this, chemical inhibitors of proteasomes can be tested for their ability to block 

cytosolic proteolysis of aldolase B. This will have to be done in the presence of 

lysosomal inhibition to detect the cytosolic protease activity while excluding lysosomal 

degradative mechanisms. It was also proposed that aldolase B might be an endogenous 

inhibitor of proteasomes. Proteasomes have been effectively purified from cells, and in 

vitro inhibition of its activity aldolase B can be tested. This study purified catalytically 

active recombinant aldolase B expressed in E. coli. The protease-sensitive carboxyl 

terminus of aldolase B would be a reasonable site for initiation of proteolysis by the 

proteasome. Carboxyl-terminally truncated recombinant aldolase B could be purified to 

determine if the "loose" carboxyl terminus is necessary for interaction with the 

proteasome. In this way, it might be shown that aldolase B is a new endogenous 

inhibitor for the proteasome or a mechanism for protein interaction with the proteasome 

might be established. 

The last major find in this study is the existence of a targeting signal in aldolase B 

required for starvation-induced degradation. Whether aldolase follows the currently 

proposed mechanism for receptor-mediated pathway is not known. This could be 

established by in vitro assays developed by J. Fred Dice using known substrates to 

compete for the pathway. Transient overexpression of the receptor for this pathway, 

LGP96, in cultured cells was shown to more than double long-lived protein degradation 

(Cuervo et al., 1996). In this study, transiently expressed aldolase B had degradation 

rates similar to those for endogenously expressed aldolase B. By coexpressing LGP96 

and aldolase B, degradation of aldolase B should be increased, and this increase should 

not occur for aldolase B with a mutated targeting signal (altered glutamine #111). This 

would demonstrate that aldolase B is a bonafide substrate for the established pathway 

and would be the first demonstration of signal function in living cells. If aldolase B 

degradation is not mediated by LGP96, then an alternative recognition mechanism or 

receptor would be indicated for aldolase B. If so, future effort would focus on 

characterizing the nature of the signal in aldolase B and identifying the novel recognition 

molecules mediating its starvation-induce degradation. 

In conclusion, ubiquitinated forms of aldolase B (e.g. Ub68) associated with 

autophagic vacuoles and lysosomes during nutrient stress (starvation) in rat liver and the 
Fao rat hepatoma cell line. Ubiquitinated aldolase B that maintained a denatured 
conformation possibly exposing recognition signals was detected in rat liver, Fao cells, 
and in the E36 Chinese hamster lung cell line. During heat stress, accumulation of 
endogenous aldolase A in lysosomes and lysosomal proteolytic processing of exogenous 
aldolase B required ubiquitination in E36 cells. Heat stress caused ubiquitin-mediated 
autophagic degradation of long-lived proteins in E36 cells. Together the results support 
a role for ubiquitination in the stress-induced degradation of aldolase B. Basal and 
starvation-induced degradation of transiently overexpressed aldolase B was like that for 
endogenous aldolase B, in the HuH7 human hepatoma cell line. A mutated (one amino 
acid changed) lysosomal targeting signal in aldolase B specifically prevented starvation- 
induced degradation of the mutant protein with no effect on other aldolase B properties. 
Together, these results supported the original hypothesis that during stress, aldolase B 
requires both ubiquitination and a receptor-mediated targeting signal for enhanced 
degradation in lysosomes. 


Alefounder, P. R., Baldwin, S. A., Perham, R. N, and Short, N. J. (1989). Cloning, 
sequence analysis and over-expression of the gene for the class II fructose 1,6- 
bisphophate aldolase of Escherichia coli. Biochem J 257, 529-534. 

Aniento, F., Roche, E., Cuervo, A. M., and Knecht, E. (1993). Uptake and degradation 
of Glyceraldehyde-3 -phosphate dehydrodenase by rat liver lysosomes. J Biol Chem 268, 

Asaka, M., Nagase, K., and Alpert, E. (1983). Chemical and clinical studies of aldolase 
isozymes in human cancer. Curr Topics Biol Med Res 11, 183-195. 

Ausubel, F. M, Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A, 
Struhl, K., Albright, L. M., and Coen, D. M. (1994). Current Protocols in Molecular 
Biology. In Current Protocols, K. Janssen, ed. (New York: Greene Publishing Associates 
Inc and John Wiley & Sons Inc), pp. variable with supplements. 

Barbas, C. F., Heine, A., Zhong, G, Hoffmann, T., Gamatikova, S., Bjomestedt, R., 
List, B., Anderson, J., Stura, E., Wilson, I. A., and Lerner, R. A. (1997). Immune Versus 
Natural Selection: Antibody Aldolases with Enzymic Rates But Broader Scope. Science 
278, 2085-2092. 

Bates, P. J., Coetzee, G. A., and VanDerWesthuyzen, D. R. (1982). The degradation of 
endogenous and exogenous proteins in cultured smooth muscle cells. Biochim Biophys 
Acta 779, 377-387. 

Beernink, P. T., and Tolan, D. R. (1992). Construction of a High-Copy "ATG Vector" 
for Expression in Escherichia coli. Protein Expression and Purification 3, 332-336. 

Berthiaume, L., Tolan, D. R., and Sygusch, J. (1993). Differential Usage of the 
Carboxyl-terminal Region among Aldolase Isozymes. J Biol Chem 268, 10826-10835. 


Beyreuther, K., Dyrks, T., Hilbich, C, Monning, U., Konig, G., Molthaup, G., Pollwein, 
P., and Masters, C. L. (1992). Amyloid Precursor Protein (APP) and beta-A4 amyloid. 
In Down Syndrome and Alzheimer Disease, Nadel and Epstein, eds. (New York: Wiley- 
Lisslnc), pp. 159-182. 

Bond, J. S., and Aronson, N. N. (1983). Endocytosis and Degradation of Native, 
Cathepsin D-Degraded, and Glutathione-Inactivated Aldolase by Perfused Rat Liver. 
Arch Biochem Biophys 227, 367-372. 

Bond, J. S. and Butler, P. E. (1987). Intracellular Proteases. Ann Rev Biochem 56:333- 

Bond, J. S., and Offermann, M. K. (1981). Intitial Events in the Degradation of Soluble 
Cellular Enzymes: Factors Affecting the Stability and Proteolytic Susceptibility of 
Fructose- 1,6-bisphosphate Aldolase. Acta Biol Med Germ 40, 1365-1374. 

Brooks, C. C, and Tolan, D. R. (1994). A partially active mutant aldolase B from a 
patient with hereditary fructose intolerance. FASEB Journal 8, 107-1 13. 

Cantor, C. R., and Schimmel, P. R. (1980). Kinetics of ligand interactions. In 
Biophysical Chemistry Part III: The behavior of biological macromolecules (New York: 
W H Freeman and Co), pp. 887-937. 

Chappel, A., Hoogenraad, N. J., and Holmes, R. S. (1978). Purification and properties of 
the native form of rabbit liver aldolase! Evidence for proteolytic modification after tissue 
extraction. Biochem J 175, 377-382. 

Chau, V., Tobias, J. W., Bachmair, A, Marriott, D., Ecker, D. J., Gonda, D. K., and 
Varshavsky, A. (1989). A Multiubiquitin Chain is Confined to Specific Lysine in a 
Targeted Short-lived Protein. Science 243, 1576-1583. 

Chiang, H.-L., Terlecky, S. R, Plant, C. P., and Dice, J. F. (1989). A Role for a 70- 
Kilodalton Heat Shock Protein in Lysosomal Degradation of Intracellular Protein. 
Science 246, 382-385. 

Chowdary, D. R., Dermody, J. J., Jha, K. K., and Ozer, H. L. (1994). Accumulation of 
p53 in a mutant cell line defective in the ubiquitin pathway. Molecular and Cellular 
Biology 14, 1997-2003. 


Chu, E. H., Brimer, P., Jacobson, K. B., and Merriam, E. V. (1969). Mammalian cell 
genetics. I. Selection and Characterization of Mutations Auxotrophic for L-glutamine 
or Resistant to 8-azaguanine in Chinese hamster cells in vitro. Genetics 62, 359-377. 

Chu, E. H. Y., and Mailing, H. V. (1968). Mammalian cell genetics. II. Chemical 
induction of of specific locus mutations in Chinese hamster cells in vitro. Proc Natl Acad 
SciUSA67, 1306-1312. 

Ciechanover, A., and Schwartz, A. (1994). The ubiquitin-mediated proteolytic pathway: 
mechanisms of recognition of the proteolytic substrate and involvement in the 
degradation of native cellular proteins. FASEB J 8, 182-191. 

Clarke, F., Stephan, P., Morton, D., and Weidemann, J. (1982). Glycolytic Enzyme 
Organization via the Cytoskeleton and its Role in Metabolic Regulation. In Regulation of 
Carbohydrate Metabolism, pp. 1-31. 

Coligan, J. E., Dunn, B. M., Ploegh, H. L., Speicher, D. W., and Wingfield, P. T. 
(1995). Current Protocols in Protein Science. In Current Protocols, V. B. Chanda, ed. 
(New York: Greene Publishing Associates Inc and John Wiley & Sons Inc), pp. variable 
with supplements. 

Cox, T. M. (1994). Aldolase B and fructose intolerance. FASEB Journal 8, 62-71. 

Cuervo, A. M., and Dice, J. F. (1996). A Receptor for the Selective Uptake and 
Degradation of Proteins by Lysosomes. Science 273, 501-503. 

Deschatrette, J., Moore, E. E., Dubois, M., Cassio, D., and Weiss, M. C. (1979). 
Dedifferentiated Variants of a Rat Hepatoma: Analysis by Cell Hybridization. Som Cell 
Gen 5, 697-718. 

Deschatrette, J., and Weiss, M. C. (1974). Characterization of differentiated and 
dedifferentiated clones from a rat hepatoma. Biochimie 56, 1603-161 1. 

Deshaies, J. R. (1995). Make it or break it: the role of ubiquitin-dependent proteolysis in 
cellular regulation. Trends Cell Biol 5, 428-434. 

Dice, J. F. (1989). Altered intracellular protein degreadation in ageing: a possible cause 
of proliferative arrest. Exp Gerontol 24, 451-459. 

Dice, J. F., and Chiang, H.-L. (1988). Peptide Signals for Protein Degradation within 
Lysosomes. Biochem Soc Symp 55, 45-55. 

Dice, J. F., and Walker, C. D. (1980). Protein degradation in metabolic and nutritional 
disorders. In Protein Degradation in Health and Disease (Ciba Foundation: Exerpta 
Medica), pp. 331-350. 

Doherty, F. J., and Mayer, R. J. (1992). The protein turnover 'cycle'. In Intracellular 
Protein Degradation (New York: IRL Press), pp. 1-8. 

Dunn, W. A. (1990). Studies on the Mechanisms of Autophagy: Formation of the 
Autophagic Vacuole. J Cell Biol 110, 1923-1933. 

Dunn, W. A. (1990). Studies on the Mechanisms of Autophagy: Maturation of the 
Autophagic Vacuole. J Cell Biol 110, 1935-1945. 

Emery, E. S., and Baldwin, R. L. (1967). Turnover of Several Mammary Enzymes 
During Lactation. Biochim Biophys Acta 136, 223-232. 

Erickson-Viitanen, S., Balestreri, E., McDermott, M., and Horecker, B. L. (1986). Sites 
of Cleavage of rabbit muscle aldolase by purified cathepsin M from rabbit liver. Biochem 
Med Met Biol 35, 191-198. 

Ford, D. K., and Yerganian, G. (1958). Observations on the chromosomes of Chinese 
hamster cells in tissue culture. Journal of the National Cancer Institute 21, 393-425. 

Freifelder, D. (1982). Application of Free Energy and Biological Processes. In Principles 
of Physical Chemistry with Applications to the Biological Sciences (Boston: Jones and 
Bartlett Publishers Inc), pp. 177-190. 

Gamblin, S. J., Davies, G. J., Grimes, J. M., Jackson, R. M., Littlechild, J. A., and 
Watson, H. C. (1991). Activity and Specificity of Human Aldolases. J of Mol Biol 219, 


Gillin, F. D., Roufa, D. J., Beoudet, A. L., and Caskey, C. T. (1972). 8-Azaguanine 
Resistance in Mammalian Cells I. Hypoxanthine-Guanine Phosphoribosyltransferase. 
Genetics 72, 239-252. 

Gorinsky, B., Laskowski, R. A., Lee, D. A., and Bomford, A. (1996). Conformational 
analysis of pentapeptide sequences matching a proposed recognition motif for lysosomal 
degradation. Biochim Biophys Acta 1293, 243-253. 

Gropper, R., Brandt, R. A., Elias, S., Bearer, C. F., Mayer, A, Schwartz, A. L., and 
Ciechanover, A. (1991). The Ubiquitin- Activating Enzyme, El, Is Required for Stress- 
Induced Lysosomal Degradation of Cellular Proteins. The J Biol Chem 266, 3602-3610. 

Guo, G. G , Gu, M., and Etlinger, J.D. (1994). 240-kDa Proteasome inhibitor (CF-2) is 
Identical to 5- Aminolevulinic Acid Dehydratase. J Biol Chem 269, 12399-12402. 

Guyton, A. C. (1979). Blood Flow Through the Systematic Circulation and Its 
Regulation. In Physiology of the Human Body (Philadelphia: Saunders College 
Publishing), pp. 138-147. 

Haimoto, H., Kurobe, N., Hosoda, S., and Kato, K. (1989). Sensitive enzyme 
immunoassay for human aldolase B. Clin Chim Acta 181, 27-36. 

Hall, N. A., Lake, B. D., Dewji, N. N., and Patrick, A. D. (1991). Lysosomal storage of 
subunit c of mitochondrial ATP synthase in Batten's disease (ceroid-lipofuscinosis). 
Biochem J 275, 269-272. 

Handley-Gearhart, P. M., Stephen, A. G., Trausch-Azar, J. S., Ciechanover, A., and 
Schwartz, A. L. (1994). Human Ubiquitin-activating Enzyme, El. J Biol Chem 269, 

Handley-Gearhart, P. M., Trausch-Azar, J. S., Ciechanover, A., and Schwartz, A L. 
(1994). Rescue of the complex temperature-sensitive phenotype of Chinese hamster 
ovary E36ts20 cells by expression of the human ubiquitin-activating enzyme cDNA. 
Biochem J 304, 1015-1020. 

Hendil, K. B., Lauridsen, A.-M. B., and Seglen, P. O. (1990). Both Endocytic and 
Endogenous Protein Degradation in Fibroblasts Is Stimulated By Serum/ Amino Acid 
Deprivation and Inhibited By 3-Methyladenine. Biochem J 272, 577-581. 


Henell, F., Berkenstam, A., Ahlberg, J., and Glaumann, H. (1987). Degradation of 
Short- and Long-Lived Proteins in Perfused Liver and in Isolated Autophagic Vacuoles- 
Lysosomes. Exp Mol Path 46, 1-14. 

Hershko, A., and Ciechanover, A. (1982). Mechanisms of Intracellular Protein 
Breakdown. Ann Rev Bioch 51, 335-364. 

Hershko, A, and Ciechanover, A. (1992). The Ubiquitin System for Protein 
Degradation. Annu Rev Bioch 61, 761-807. 

Hirschberg, J., and Marcus, M. (1982). Isolation by a replica-plating technique of 
Chinese hamster temperature-sensitive cell cycle mutants. J Cellular Phys 113, 159-166. 

Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, J. K., and Pease, L. R. (1989). Site- 
directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77, 

Hochstrasser, M. (1996). Protein Degradation or Regulation: Ub the Judge. Cell 84, 

Hochstrasser, M. (1992). Ubiquitin and intracellular protein degradation. Curr Opin Cell 
BioW, 1024-1031. 

Hopgood, M. F., Knowles, S. E., Bond, J. S., and Ballard, F. J. (1988). Degradation of 
native and modified forms of fructose-bisphosphate aldolase microinjected into HeLa 
cells. Biochem J 256, 81-88. 

Hough, R. and Rechsteiner, M. (1984). Effects of temperature on the degradation of 
proteins in rabbit reticulocyte lysates and after injection into HeLa cells. Proc Natl Acad 
Sci USA 81, 90-94. 

Horecker, B. L., Erickson-Viitanen, S., Melloni, E., and Pontremoli, S. (1985). 
Inactivation of Rabbit Liver and Muscle Aldolases by Limited Proteolysis by Lysosomal 
Cathepsin M. Curr Top Cell Regul 25, 77-89. 

Hough, R., Pratt, G., and Rechsteiner, M. (1986). Ubiquitin-lysozyme conjugates. 
Identification and characterization of an ATP-dependent protease from rabbit 
reticulocyte lysates. J Biol Chem 261, 2400-2408. 

Jentsch, S. (1992). The Ubiquitin-conjugation System. Annu Rev Genet 26, 179-207. 

Kelly, J. H., and Darlington, G. J. (1989). Modulation of the Liver Specific Phenotype in 
the Human Hepatoblasoma Line HepG2. In Vitro Cell & Dev Biol 25, 217-222. 

Knowles, S. E., Hopgood, M. F., and Ballard, F. J. (1989). Effects of Inhibitors on 
Aldolase Breakdown After its Microinjection into HeLa Cells. Biochem J 259, 27-33. 

Kominami, E., Hashida, S., Khairallah, E. A., and Katunuma, N. (1983). Sequestration 
of Cytoplasmic Enzymes in an Autophagic Vacuole-Lysosomal System Induced by 
Injection of Leupeptin. J Biol Chem 258, 6093-6100. 

Kopitz, J., Kisen, G. O., Gordon, P. B., Bohley, P., and Seglen, P. O. (1990). Non- 
selective autophagy of cytosolic enzymes by isolated rat hepatocytes. J Cell Biol 111, 

Kuehl, L., and Sumsion, E. N. (1970). Turnover of Several Glycolytic Enzymes in Rat 
Liver. J Biol Chem 245, 6616-6623. 

Kulka, R. G., Raboy, B., Schuster, R., Parag, H. A., Diamond, G., Ciechonover, A., and 
Marcus, M. (1988). A Chinese Hamster Cell Cycle Mutant Arrested at G2 Phase Has a 
Temperature- Sensitive Ubiquitin-Activating Enzyme, El. J Biol Chem 263, 15726- 

Kusakabe, T., Motoki, K., and Hori, K. (1997). Mode of Interactions of Human 
Aldolase Isozymes with Cytoskeletons. Arch Biochem Biophys 344, 184-193. 

Lawrence, B. P., and Brown, W. J. (1992). Autophagic vacuoles rapidly fuse with pre- 
existing lysosomes in cultured hepatocytes. J Cell Sci 102, 515-526. 

Lebherz, H, and Rutter, W. (1969). Distribution of fructose diphosphate aldolase 
variants in biological systems. Biochemistry 8, 109-121. 

Lebherz, H. G. (1972). Stability of quaternary structure of mammalian and avian fructose 
diphosphate aldolases. Biochemistry 11, 2243-2250. 

Lebherz, H. G. (1975). Evidence for the Lack of Subunit Exchange between Aldolase 
Tetramers m Vivo. J Biol Chem 250, 7388-7391. 

Lefer, A. M, Messenger, M., and Okamatsu, S. (1982). Salutary Actions of 
Thromboxane Synthetase Inhibition During Global Myocardial Ischemia. Nouyn- 
Schmiedeberg's Arch Pharmacol 321, 130-134. 

Lenk, S. E., Susan, P. P., Hickson, I., Jasionowski, T., and Dunn, W. A.Jr. (Submitted 
1998). Ubiquitinated Aldolase Accumulates in Stress-Induced Autophagic Vacuoles. 

Lenk, S. L., Dunn, W. A. Jr., Trausch, J. S., Ciechanover, A., and Schwartz, A. L. 
(1992). Ubiquitin-activating Enzyme, El, is Associated with Maturation of Autophagic 
Vacuoles. J Cell Biol 118, 301-308. 

LLovera, M., Garcia-Martinez, C, Agell, N., Lopez- Soriano, F. J., and Argiles, J. M. 
(1995). Muscle wasting associated with cancer cachexia is linked to an important 
activation of the ATP-dependent ubiquitin-mediated proteolysis. International J Cancer 
61, 138-141. 

Lowe, J., Stock, D., Jap, B., Zwicki, P., Baumeister, W., and Huber, R. (1995). Crystal 
Structure of the 20S Proteasome from the Archaeon T. acidophilum at 3.4 A Resolution. 
Science 268, 533-539. 

Marchand, M., Pliszczak, A., Gibson, W. C, Wierenga, R. K., Opperdoes, F. R., and 
Michels, P. A. M. (1988). Characterization of the genes for fructose-bisphosphate 
aldolase in Trypanosoma brucei. Mol Biochem Parasit 29, 65-76. 

Mortimore, G. E., and Kadowaki, M. (1994). Autophagy: Its Mechanism and 
Regulation. In Cellular Proteolytic Systems, A. J. Ciechanover and A. L. Schwartz, eds. 
(New York: Wiley-Liss). 

Mortimore, G. E., and Poso, A. R (1987). Intracellular protein catabolism and its 
control during nutrient deprivation and supply. Ann Rev Nutr 7, 539-564. 

Nakabayashi, H., Taketa, K., Miyano, K., Yamane, T., and Sato, J. (1982). Growth of 
Human Hepatoma Cell Lines with Differentiated Functions in Chemically Defined 
Medium. Cancer Res 42, 3858-3863. 

Offermann, M. K., Chlebowski, J. F., and Bond, J. S. (1983). Action of cathepsin D on 
fructose- 1,6-bisphosphate aldolase. Biochem J. 211, 529-534. 

Olson, T. S., Terlecky, S. R., and Dice, J. F. (1990). Pathways of Intracellular Protein 
Degradation in Eukaryotic Cells. In Pharmaceutical Biotechnology: Stability of Protein 
Pharmaceuticals and Strategies for Stabilization, T. J. Ahearn and M. C. Manning, eds. 
(New York: Plenum Press). 

O'Reilly, G., and Clarke, F. (1993). Identification of an actin binding region in aldolase. 
FEBS 321, 69-72. 

Paolella, G., Santamaria, T., Izzo, P., Costanzo, P., and Salvatore, F. (1984). Isolation 
and Nucleic Acid Sequence of a Full-Length cDNA Coding for Aldolase B from Human 
Liver. Nucleic Acids Res 12, 7401-7410. 

Penhoet, E. E., and Rutter, W. J. (1975). Detection and Isolation of Mammalian 
Fructose-Diphosphate Aldolases. Meth Enzymol 42, 240-258. 

Pontremoli, S., and Melloni, E. (1986). Extralysosomal Protein Degradation. Ann Rev 
Bioch 55, 455-481. 

Pontremoli, S., Melloni, E., Damiani, G, Michetti, M., Salamino, F., Sparatore, B., and 
Horecker, B. L. (1984). Binding of Monoclonal Antibody to Cathepsin M Located on 
the External Surface of Rabbit Lysosomes. Arch Biochem Biophys 233, 267-271. 

Pontremoli, S., Melloni, E., Michetti, M., Salamino, F., Sparatore, B., and Horecker, B. 
L. (1982). Characterization of the inactive form of fructose- 1,6-bisphosphate aldolase 
isolated from livers of fasted rabbits. Proc Natl Acad Sci USA 79, 5194-7196. 

Pontremoli, S., Melloni, E., Salamino, F., Sparatore, B., Michetti, M., and Horecker, B. 
L. (1982). Cathepsin M: a lysosomal proteinase with aldolase-inactivating activity. Arch 
Biochem Biophys 214, 376-385. 

Pontremoli, S., Melloni, E., Salamino, F., Sparatore, B., Michetti, M, and Horecker, B. 
L. (1979). Changes in activity of fructose- 1,6-bisphosphate aldolase in livers of fasted 
rabbits and accumulation of crossreacting immune material. Proc Natl Acad Sci USA 76, 

Prusiner, S. B., and DeArmond, S. J. (1987). Prions causing nervous system damage. 
Lab Invest 56, 349-363. 

Reuber, M. D. (1961). A Transplantable Bile-Secreting Hepatocellular Carcinoma in the 
Rat. J Natl Cancer Inst 26, 891-897. 

Reznick, A. Z., Rosenfelder, L., Shpund, S., and Gershon, D. (1985). Identification of 
Intracellular Degradation Intermediates of Aldolase B By Antiserum to the Denatured 
En2yme. Proc Natl Acad Sci USA 82, 6114-8118. 

Rickwood, D. (1992). Preparative Centrifugation: A Practical Approach. In The 
Practical Approach Series, D. Rickwood and B. D. Hames, eds. (New York: IRL Press). 

Rottmann, W. H., Tolan, D. R, and Penhoet, E. E. (1984). Complete Amino Acid 
Sequence for Human Aldolase B Derived From cDNA and Genomic Clones. Proc Natl 
Acad Sci USA 81, 2738-2742. 

Rock, K.L., Gramm, C, Rothstein, L., Clark, K., Stein, R., Dick, L., Hwang, D., 
Goldberg, A.L. (1994). Inhibitors of the proteasome Block the degradation of most cell 
proteins and the generation of peptides presented on MHC class I molecules. Cell 78, 

Ruderman, N. B. (1975). Muscle amino acid metabolism and gluconeogenesis. Ann Rev 
Med 26, 245-266. 

Rutter, W. J., Hunsley, J. R., Groves, W. E., Calder, J., Rajkumar, T. V., and Woodfin, 
B. M. (1966). Fructose Diphosphate Aldolase. Meth Enzymol 9, 479-498. 

Sakakibara, M., Takahashi, I., Takasaki, Y., Mukai, T., and Hori, K. (1989). 
Construction and Expression of Human aldolase A and B Expression Plasmids in 
Escherichia coli Host. Biochim Biophys Acta 1007, 334-342. 

Schwartz, A. L , Brandt, R. A., Geuze, H, and Ciechanover, A. (1992). Stress-induced 
alterations in autophagic pathway: relationship to ubiquitin system. Am J Phys 262, 

Seglen, P. O. and Bohley, P. (1992). Autophagy and other vacuolar protein degradation 
mechanisms. Experiential, 158-172. 

Seglen, P. O., and Gordon, P. B. (1982). 3-Methyladenine: Specific Inhibitor of 
Autophagic/Lysosomal Protein Degradation in Isolated Rat Hepatocytes. Proc Natl 
Acad Sci USA 79, 1889-1892. 

Stein, L. B., and Arias, I. M. (1976). Physiology of the Liver. In Gastroenterology, H. L. 
Bockus, J. E. Berk, W. S. Haubrich, M. Kaiser, J. L. A. Roth and F. Vilardell, eds. 
(Philadelphia: W B Saunders Co), pp. 40-67. 

Susan, P. P., Lenk, S. E., Hickson, I., and Dunn, W. A. (1995). Multi-Ubiquitination of 
a Cytosolic Protein Destined for Lysosomal Degradation. Mol Biol Cell 6, 105a. 

Sygusch, J., Beaudry, D., and Allaire, M. (1990). Inactivation of Mammalian Fructose 
Diphosphate Aldolases by COOH Terminus Autophosphorylation. Arch Bioch Biophys 

Takahashi, I., Takasaki, Y., and Hori, K. (1989). Site-Directed Mutagenesis of Human 
Aldolase Isozymes: The Roe of Cys-72 and Cys-338 Residues of Aldolase A and the 
Carboxy-Terminal Tyr Residues of Aldolases A and B. J Biochem J 05, 281-286. 

Terlecky, S. R., Chiang, H.-L., Olson, T. S., and Dice, J. F. (1992). Protein and Peptide 
Binding and Stimulation of in Vitro Lysosomal Proteolysis by the 73-kDa Heat Shock 
Cognate Protein. J Biol Chem 267, 9202-9209. 

Trausch, J. S., Grenfell, S. J., Handley-Gearhart, P. M., Ciechanover, A., and Schwartz, 
A. L. (1993). Immunofluorescent localization of the ubiquitin-activating enzyme, El, to 
the nucleus and cytoskeleton. Am J Phys 264, C93-C102. 

Tsutsumi, K., Mukai, T., Tsutsumi, T., Mori, M , Daimon, M., Tanaka, T., Yatsuki, H, 
Hori, K., and Ishikawa, K. (1984). Nucleotide Sequence of Rat Liver Aldolase B 
Messenger RNA J Biol Chem 259, 14572-14575. 

Tuttle, D. L., Lewin, A. S., and Dunn, W. A. (1993). Selective Autophagy of 
Peroxisomes in Methylotrophic Yeast. Eur J Cell Biol 60, 283-290. 

Ueno, T., and Kominami, E. (1991). Mechanism and regulation of lysosomal 
sequestration and proteolysis. Biomed Biochim Acta 50, 365-371. 

Ueno, T., Watanabe, S., Hirose, M, Namihisa, T., and Kominami, E. (1990). Phalloidin- 
Induced Accumulation of Myosin in Rat Hepatocytes is Caused by Suppression of 
Autolysosome Formation. European J Biochem 190, 63-69. 

Verlessy, B. G., and Vas, M. (1992). Metabolite Channeling Versus Free Diffusion: 
Reinterpretation of Aldolase-Catalyzed Inactivation of Glyceraldehyde-3 -phosphate 
Dehydrogenase. Biochem J 282, 977-979. 

Wang, J., Morris, A. J., Tolan, D. R, and Pagliaro, L. (1996). The Molecular Nature of 
the F-actin Binding Activity of Aldolase Revealed with Site-directed Mutants. J Biol 
Chem 271, 6861-6865. 

Waxman, L., Fagan, J. M., and Goldberg, A. L. (1987). Demonstration of Two Distinct 
High Molecular Weight Proteases in Rabbit Reticulocytes, One of Which Degrades 
Ubiquitin Conjugates. J Biol Chem 262, 2451-2457. 

Wing, S. S., Chiang, H.-L., Goldberg, A. L., and Dice, J. F. (1991). Proteins containing 
peptide sequences related to Lys-Phe-Glu-Arg-Gln are selectively depleted in liver and 
heart but not skeletal muscle of fasted rats. Biochem J 275, 165-169. 

Zhang, R., Yatsuki, H., Kusakabe, T., Iwabe, N, Miyata, T., Imai, T., Yoshida, M., and 
Hori, K. (1995). Structures of cDNAs Encoding the Muscle-Type and Non-Muscle- 
Type Isozymes of Lamprey Fructose Bisphosphate Aldolases and the Evolution of 
Aldolase Genes. J Biochem 117, 545-553. 

In 1986, Peter P. Susan earned his Bachelor of Science degree from the 
Pennsylvania State University at State College, Pennsylvania. Early in his studies he 
majored in computer science then changed his major to microbiology. When finishing his 
undergraduate work, Mr. Susan moved to Gainesville, Florida where he enrolled in the 
Vertebrate Zoology Program of the College of Liberal Arts and Sciences at the 
University of Florida and worked full-time as a laboratory technician in the Departments 
of Zoology, Botany, and Biological Sciences, providing support for research and 
teaching. During this work, he prepared technical illustrations published in laboratory 
manuals for the University of Florida. 

In 1988, Peter Susan then successfully completed a Masters Fellowship for 
Teachers in the Department of Instruction and Curriculum of the College of Education at 
the University of Florida, earning a master of education degree in science instruction and 
curriculum. For several years, he served as a public school science teacher in Manatee 
County, Florida. 

By spring of 1 992, Mr. Susan pursued and acquired a research assistantship in 
the Department of Anatomy and Cell Biology of the College of Medicine at the 
University of Florida. In 1993, he began studying protein turmover in the laboratory of 
Dr. William A. Dunn, Jr. This dissertation resulted from much of that work. 


I certify that I have read this study and that in my opinion it conforms to acceptable 
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for 
the degree of Doctor of Philosophy. 

William A. Dunn, Jr., Chair ' 
Associate Professor of Anatomy and 
Cell Biology 

I certify that I have read this study and that in my opinion it conforms to acceptable 
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for 
the degree of Doctor of Philosophy. 

Brian D. Cain 

Associate Professor of Biochemistry 
and Molecular Biology 

I certify that I have read this study and that in my opinion it conforms to acceptable 
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for 
the degree of Doctor of Philosophy. 

%fUuK f**^ +- 

Mohan K. Raizada 
Processor of Physiology 

I certify that I have read this study and that in my opinion it conforms to acceptable 
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for 
the degree of Doctor of Philosophy. 

Carl M. Feldherr 
Professor of Anatomy and Cell 

This dissertation was submitted to the Graduate Faculty of the College of Medicine and to 
the Graduate School and was accepted as partial fulfillment of the requirements for the degree of 
Doctor of Philosophy. 

August 1998 

Dean, College of Medicine" 

Dean, Graduate School 



3 1262 08555 3039